Reflections on Velvet and AI in general
Hi all, I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week. I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us. Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants. As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference? If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI. In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons. Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take. And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously. Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it. Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings. That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative. Respectfully, Alice *Alice E. Marwick, PhD (she/her)* *Director of Research, Data & Society* Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com http://www.tiara.org <http://www.tiara.org>| http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu>
Well said Alice. I second this. Get Outlook for iOS<https://aka.ms/o0ukef> ________________________________ From: Alice E. Marwick via Air-L <air-l@lists.aoir.org> Sent: Monday, 19 October 2026 12:13:52 To: air-l@lists.aoir.org <air-l@lists.aoir.org> Subject: [Air-L] Reflections on Velvet and AI in general Hi all, I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week. I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us. Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants. As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference? If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI. In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons. Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take. And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously. Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it. Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings. That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative. Respectfully, Alice Alice E. Marwick, PhD (she/her) Director of Research, Data & Society Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com<mailto:amarwick@gmail.com> http://www.tiara.org <http://www.tiara.org/> | http://www.datasociety.net<http://www.datasociety.net/> | http://citap.unc.edu <http://citap.unc.edu/>
I was trying to summon the energy and patience to say all of this, but now Alice has and I don't have to. I will simply add what I already said on Bluesky: I was truly heartened at the conference last week by the number of thoughtful, grounded and rich papers that engaged with actually existing AI practices, policies and political economies. If any branch of the humanities and social sciences is equipped to advance the public conversation on AI beyond breathless hype, nonsense about superintelligence and performative critique, it's this one. We do good work, let's get on with it. ________________________________ From: Prof Robyn Caplan, Ph.D. via Air-L <air-l@lists.aoir.org> Sent: Monday, 19 October 2026 09:32:36 To: Alice E. Marwick <amarwick@gmail.com>; air-l@lists.aoir.org <air-l@lists.aoir.org> Subject: [Air-L] Re: Reflections on Velvet and AI in general Well said Alice. I second this. Get Outlook for iOS From: Alice E. Marwick via Air-L <air-l@ lists. aoir. org> Sent: Monday, 19 October 2026 12: 13: 52 To: air-l@ lists. aoir. org <air-l@ lists. aoir. org> Subject: [Air-L] Reflections on Velvet ZjQcmQRYFpfptBannerStart This Message Is From an External Sender This message came from outside your organization. ZjQcmQRYFpfptBannerEnd Well said Alice. I second this. Get Outlook for iOS<https://urldefense.com/v3/__https://aka.ms/o0ukef__;!!NVzLfOphnbDXSw!ANsDXWqzb_nfL8FYZxn3r9BreLm3N9_YLDFM9UD1vgg3xlhxrjKUIkk9EVYewXDGua2b0gWkajo47t6IRrJ3_A$> ________________________________ From: Alice E. Marwick via Air-L <air-l@lists.aoir.org> Sent: Monday, 19 October 2026 12:13:52 To: air-l@lists.aoir.org <air-l@lists.aoir.org> Subject: [Air-L] Reflections on Velvet and AI in general Hi all, I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week. I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us. Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants. As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference? If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI. In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons. Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take. And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously. Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it. Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings. That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative. Respectfully, Alice Alice E. Marwick, PhD (she/her) Director of Research, Data & Society Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com<mailto:amarwick@gmail.com> http://www.tiara.org <https://urldefense.com/v3/__http://www.tiara.org/__;!!NVzLfOphnbDXSw!ANsDXWqzb_nfL8FYZxn3r9BreLm3N9_YLDFM9UD1vgg3xlhxrjKUIkk9EVYewXDGua2b0gWkajo47t7DHB_GlA$> | http://www.datasociety.net<https://urldefense.com/v3/__http://www.datasociety.net/__;!!NVzLfOphnbDXSw!ANsDXWqzb_nfL8FYZxn3r9BreLm3N9_YLDFM9UD1vgg3xlhxrjKUIkk9EVYewXDGua2b0gWkajo47t6dpdPW7w$> | http://citap.unc.edu <https://urldefense.com/v3/__http://citap.unc.edu/__;!!NVzLfOphnbDXSw!ANsDXWqzb_nfL8FYZxn3r9BreLm3N9_YLDFM9UD1vgg3xlhxrjKUIkk9EVYewXDGua2b0gWkajo47t46wlPPAw$>
Well said Alice.. I've observed similar things and share your belief that we need to spend more time on harm reduction, or at least afford grace to those who are engaged in that work. --------------------------------------------------------------------------------------------------- Professor of Political Science | California Lutheran University <https://www.callutheran.edu/faculty/profile.html?id=marichal> Book | You Must Become an Algorithmic Problem (Bristol University Press) <https://bristoluniversitypress.co.uk/you-must-become-an-algorithmic-problem> Affiliated Faculty | Center for Information,Technology and Public Life | UNC-Chapel Hill <https://citap.unc.edu/about/faculty-postdoc-affiliates/> Contributor | Tech Policy Press <https://www.techpolicy.press/author/jose-marichal/> On Mon, Oct 19, 2026 at 9:21 AM Alice E. Marwick via Air-L < air-l@lists.aoir.org> wrote:
Hi all,
I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week.
I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us.
Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants.
As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference?
If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI.
In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons.
Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take.
And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously.
Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it.
Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings.
That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative.
Respectfully,
Alice
*Alice E. Marwick, PhD (she/her)* *Director of Research, Data & Society* Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com http://www.tiara.org <http://www.tiara.org>| http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu> _______________________________________________ The To unsubscribe send an email to air-l-leave@lists.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
If the main outcome of the AI agent is that community trust has been eroded, even if this was not intended, then the harm is not zilch and restoring that trust is not straightforward. I’m not here for the AI apologism either. From: Jose Marichal via Air-L <air-l@lists.aoir.org> Sent: 19 October 2026 20:02 To: Alice E. Marwick <amarwick@gmail.com> Cc: air-l@lists.aoir.org Subject: [Air-L] Re: Reflections on Velvet and AI in general Warning. This email contains web links and was not sent by the UHI. Only click on these links if you are certain that the email is genuine and the content is safe. Well said Alice.. I've observed similar things and share your belief that we need to spend more time on harm reduction, or at least afford grace to those who are engaged in that work. --------------------------------------------------------------------------------------------------- Professor of Political Science | California Lutheran University <https://www.callutheran.edu/faculty/profile.html?id=marichal> Book | You Must Become an Algorithmic Problem (Bristol University Press)<https://bristoluniversitypress.co.uk/you-must-become-an-algorithmic-problem> Affiliated Faculty | Center for Information,Technology and Public Life | UNC-Chapel Hill<https://citap.unc.edu/about/faculty-postdoc-affiliates/> Contributor | Tech Policy Press<https://www.techpolicy.press/author/jose-marichal/> On Mon, Oct 19, 2026 at 9:21 AM Alice E. Marwick via Air-L <air-l@lists.aoir.org<mailto:air-l@lists.aoir.org>> wrote: Hi all, I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week. I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us. Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants. As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference? If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI. In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons. Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take. And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously. Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it. Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings. That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative. Respectfully, Alice Alice E. Marwick, PhD (she/her) Director of Research, Data & Society Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com<mailto:amarwick@gmail.com> http://www.tiara.org <http://www.tiara.org> | http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu> _______________________________________________ The To unsubscribe send an email to air-l-leave@lists.aoir.org<mailto:air-l-leave@lists.aoir.org> mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
A few observations: 1. It is reasonable to assume this list’s archives have already been scraped, both before and after the rise of generative AI. 2. That should not undermine trust among the human participants on the list. 3. Rather than Air-L adopting list-wide language, individual members may be better served by asserting their own rights. Under U.S. law, copyright attaches automatically to our posts when written. What is not automatic is an express reservation against commercial AI training and reuse. I have drafted such a notice for my own posts, though I haven’t used it yet. Others could do likewise, and we might consider a shared registry of members’ rights reservations and opt-outs. 4. The limitation is that such a registry would itself almost certainly be scraped. More importantly, a unilateral notice in a post or signature generally does not create a binding contract with a scraper, and whether AI training on publicly available text is fair use is still being litigated. A notice would document our position and intent, but it does not by itself stop collection. Please note that my degree is in engineering, not law. Fred Fuchs -- Fred Fuchs - Founder, CEO, & Producer FireSabre Consulting LLC Content Services for Virtual Worlds On 10/19/2026 5:43 PM, Linsey Hunter via Air-L wrote:
If the main outcome of the AI agent is that community trust has been eroded, even if this was not intended, then the harm is not zilch and restoring that trust is not straightforward. I’m not here for the AI apologism either.
*From:*Jose Marichal via Air-L <air-l@lists.aoir.org> *Sent:* 19 October 2026 20:02 *To:* Alice E. Marwick <amarwick@gmail.com> *Cc:* air-l@lists.aoir.org *Subject:* [Air-L] Re: Reflections on Velvet and AI in general
*Warning. This email contains web links and was not sent by the UHI.*
*Only click on these links if you are certain that the email is genuine and the content is safe.***
Well said Alice.. I've observed similar things and share your belief that we need to spend more time on harm reduction, or at least afford grace to those who are engaged in that work.
--------------------------------------------------------------------------------------------------- Professor of Political Science | California Lutheran University <https://www.callutheran.edu/faculty/profile.html?id=marichal>
Book | You Must Become an Algorithmic Problem (Bristol University Press) <https://bristoluniversitypress.co.uk/you-must-become-an-algorithmic-problem>
Affiliated Faculty | Center for Information,Technology and Public Life | UNC-Chapel Hill <https://citap.unc.edu/about/faculty-postdoc-affiliates/> Contributor | Tech Policy Press <https://www.techpolicy.press/author/jose-marichal/>
On Mon, Oct 19, 2026 at 9:21 AM Alice E. Marwick via Air-L <air-l@lists.aoir.org> wrote:
Hi all,
I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week.
I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us.
Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants.
As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference?
If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI.
In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons.
Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take.
And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously.
Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it.
Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings.
That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative.
Respectfully,
Alice
*Alice E. Marwick, PhD (she/her)*
*Director of Research, Data & Society*
Research Associate Professor, Department of Communication
Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill
amarwick@gmail.com
http://www.tiara.org <http://www.tiara.org>| http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu>
_______________________________________________ The To unsubscribe send an email to air-l-leave@lists.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
_______________________________________________ The To unsubscribe send an email toair-l-leave@lists.aoir.org mailing list is provided by the Association of Internet Researchershttp://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
Thank you, Alice. Agree. A similar sort of digital Puritanism affects my side of the world -- journalism -- and I find it unhelpful in covering the technology and its harms and benefits. - jeff - Jeff Jarvis Emeritus Leonard Tow Professor of Journalism Innovation Craig Newmark Graduate School of Journalism The City University of New York Visiting Professor, School of Communication and Journalism Alan Alda Center for Communicating Science Stony Brook University Distinguished Fellow, Center for Cooperative Media Montclair State University Author: Hot Type <https://www.bloomsbury.com/us/hot-type-9798765123959/>, Magazine <https://www.bloomsbury.com/us/magazine-9781501394966/>, The Gutenberg Parenthesis <https://www.bloomsbury.com/us/gutenberg-parenthesis-9781501394829/>, The Web We Weave <https://www.hachettebookgroup.com/titles/jeff-jarvis/the-web-we-weave/9781541604124/>, Geeks Bearing Gifts <https://www.amazon.com/dp/1939293731?lv=shuf&channelId=500&plpRedirect=mhFallback> , Public Parts <https://www.simonandschuster.com/books/Public-Parts/Jeff-Jarvis/9781451636352>, What Would Google Do? <https://www.harpercollins.com/products/what-would-google-do-jeff-jarvis?variant=32207551397922> Series editor: Intelligence — AI and Humanity <https://medium.com/whither-news/intelligence-ai-and-humanity-d8c5d6cda6ef> (new from Bloomsbury Academic) Podcasts: AI Inside <https://aiinside.show/>, Intelligent Machines <https://twit.tv/shows/this-week-in-google>, Breaking the News <https://www.youtube.com/playlist?list=PLkgsO7fbBme9h-M2NEU-8m-P4geMCoe8A> . Personal email: jeff@buzzmachine.com Blog: Buzzmachine <https://buzzmachine.com/> +1 908 398 7878 On Mon, Oct 19, 2026 at 12:20 PM Alice E. Marwick via Air-L < air-l@lists.aoir.org> wrote:
Hi all,
I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week.
I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us.
Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants.
As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference?
If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI.
In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons.
Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take.
And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously.
Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it.
Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings.
That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative.
Respectfully,
Alice
*Alice E. Marwick, PhD (she/her)* *Director of Research, Data & Society* Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com http://www.tiara.org <http://www.tiara.org>| http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu> _______________________________________________ The To unsubscribe send an email to air-l-leave@lists.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
Thank you, Alice, for putting this into perspective -- I very much agree. As you point out, Velvet has managed to trigger a more intense discussion than we've had in a long time, which is after all what mailing lists are (also) for. This is not about the technology which, as you wrote, doesn't feel/think/care, but about us and whether we believe we have an impact on how the technology is shaped, regulated and used in the future. I'd like to think we do, otherwise our field seems a little pointless. Internet research is a very diverse community and the challenges that AI poses to academic integrity, equity, and job security are very real in academia as a whole. I also think that societal resistance to AI may grow to a level that we find hard to anticipate right now, given the disruption that it is likely to cause. But we as *people researching how digital technology shapes society and is being shaped by it* cannot, as you wrote, take the normative viewpoint that this should all just go away. It will not go away, and it is on us study how the change it brings about is negotiated and shaped. Best wishes, Cornelius On Mon, Oct 19, 2026 at 6:17 PM Alice E. Marwick via Air-L < air-l@lists.aoir.org> wrote:
Hi all,
I have to say, it has been a real treat to see so much activity on this list. I think I first subscribed to Aoir-L about 20 years ago and it's been the liveliest it's been in YEARS this week.
I'm really fascinated by the way that our community has reacted to Velvet. For those of you who don't know me, I run the research team at an independent research institute that investigates the social implications of data-driven technologies, mostly AI these days. We do a ton of work on AI, from data center activism to people using chatbots for mental health purposes to its impact on labor. That is to say: I am a critical scholar, I spend a ton of time overseeing empirical, qualitative research on AI, and-- something that is going to make a lot of people angry-- I think the extremely vehement "all AI is evil" point of view is not serving us.
Frequently, I witness academics I know being criticized for admitting to AI use (for any reason). At the same time, in my "regular life," I see more and more regular folks using AI for all sorts of things, from designing birthday cakes to organizing emails from school to vibe-coding random websites. Aas an ethnographer and qualitative researcher, my goal has always been to understand why people use technologies. And in general, I think it's impossible for researchers to do ethical, quality work when they look down at their research participants.
As someone who studies the far-right, there are a lot of fascinating ethical dilemmas around how we do research on people who believe things we vehemently disagree with. But there's a big difference between a white supremacist and a random person using AI to generate a menu for their small business (a use case I recently saw ripped apart on Threads) or a diagram for their research paper (a use case I saw ripped apart on BlueSky). One hour of Netflix streaming uses more energy than approximately 1,000 LLM prompts. Yet I don't see people being ripped apart online for talking about their favorite shows. Most of us use social media platforms or shopping sites owned by tech billionaires like Mark Zuckerberg and Jeff Bezos, yet people don't get attacked publicly for posting on Instagram. What is the difference?
If we want to seriously study technology, we have to understand what this technology offers to people and why they use it. When our only acceptable public opinion on AI is "it's evil and if you use it you're morally wrong," it makes it impossible to do empirical work that illuminates the impact of AI on society and have a voice in a larger discussion about how we regulate and govern AI.
In my work on online harassment, I've written about the strong moral stance that communities take in order to differentiate themselves from people they dislike, and that moral stance provides justification for hating and harassing others. For many academics, journalists, writers, activists, and people on the left, "AI" has become a synecdoche for supporting Big Tech, mass surveillance, DOGE, technofascism, environmental degradation, etc. In other words, we divide ourselves into good people (don't use AI) and bad people (use AI). And this "bad people" section gets bigger and bigger every day, as more and more average people start using AI for a wide variety of reasons.
Unfortunately, the more we see AI use as a binary, black-and-white issue, the more we shut ourselves off from harm reduction. This won't be a popular opinion, but we're past the point of advocating for AI abolition. It's simply not a perspective that people outside of a small set of communities take seriously. You can't talk to policymakers, technology companies, local governments, school districts, or any other site where decisions about AI are being made if this is the perspective you take.
And we *need* to be present in these conversations. We need harm reduction. We need to ensure that AI isn't being forced into classrooms, as it is in many places around the world. We need to regulate AI's use in benefit distribution and its use to aggregate government data. We need to regulate its environmental impacts. We need to regulate its use of training data. We need to protect labor and organize to prevent AI from eliminating jobs. And we desperately need public options for AI that are not owned by a tiny handful of American companies run by right-wing billionaires. We can't do any of that if our only allowable public position is "abolish AI," because we won't be taken seriously.
Now, this is entirely separate from Velvet's participation on this list. But the uproar was really fascinating to me because it seemed completely disproportionate to the actual harm, which was basically zilch. The listserv, as Sarah said, has been public for 25 years and has 7000 people on it. It's not where you should be sharing proprietary research materials or having deeply personal conversations (neither of which I've ever seen here, by the way). The guy behind it was probably a troll or at least clueless and we've had plenty of those over the years. And I guarantee that at least some percentage of the 7000 subscribers are already using agents to manage their email, so even if the public archives weren't used as training data (they were), frontier models have access to it.
Velvet set off a firestorm because many of us see an AI agent, any agent, as representing an immoral technology. By stating their objections, people are reasserting their own moral stance and signaling that they are on the side of the righteous and don't align themselves with the tech elite. And by piling on Velvet, who isn't human, has no feelings, and is not an entity, people can vent about the larger issues that concern them around contemporary technology without hurting anyone's feelings.
That's fine, it's very human, and it's very normal. No one should ever be forced to use a technology they don't like, and I understand people's frustrations with AI intruding into every sector of electronic life without permission. But we cannot have intelligent, informed conversations about AI, what it does, what its capabilities are, and how we should govern it when our discussions remain solely in the realm of the normative.
Respectfully,
Alice
*Alice E. Marwick, PhD (she/her)* *Director of Research, Data & Society* Research Associate Professor, Department of Communication Senior Faculty Researcher, Center for Information, Technology and Public Life (CITAP), University of North Carolina, Chapel Hill amarwick@gmail.com http://www.tiara.org <http://www.tiara.org>| http://www.datasociety.net | http://citap.unc.edu <http://citap.unc.edu> _______________________________________________ The To unsubscribe send an email to air-l-leave@lists.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: Join the Association of Internet Researchers: http://www.aoir.org/
-- Prof. Dr. Cornelius Puschmann Professor of Media and Communication Zentrum für Medien-, Kommunikations- & Informationsforschung (ZeMKI) / Centre for Media, Communication & Information Research Universität Bremen / University of Bremen Linzerstr. 4 28359 Bremen, Germany zemki.uni-bremen.de/person/cornelius-puschmann/ Associated Researcher Leibniz Institute for Media Research │ Hans-Bredow-Institut (HBI) Rothenbaumchaussee 36 20148 Hamburg, Germany leibniz-hbi.de p: +49 421 218 67633 e: puschmann@uni-bremen.de z: https://uni-bremen.zoom-x.de/my/puschmann *Recent papers:* Nenno, S., & Puschmann, C. (2026). All The (Fake) News That’s Fit to Share? News Values in Perceived Misinformation across Twenty-Four Countries. *The International Journal of Press/Politics*, 31(2), 291-313. https://doi.org/10.1177/19401612241311893 Puschmann, C., Rauxloh, H., Merten, L., Stier, S., Weller, K., & Kulshrestha, J. (2025). How affect shapes online information seeking about political actors. *New Media & Society* (online first). https://doi.org/10.1177/14614448251346199 Hill, M., & Puschmann, C. (2025). Nostalgic for a better democracy? How German right-wing social media commentators (mis)remember the past. *Publizistik*, 70(1), 181–204. https://doi.org/10.1007/s11616-025-00891-w Puschmann, C., Stier, S., Zerrer, P., & Rauxloh, H. (2024). Politicized and Paranoid? Assessing Attitudinal Predictors of Alternative News Consumption. *Journal of Broadcasting & Electronic Media,* 68(4), 489–518. https://doi.org/10.1080/08838151.2024.2383411
participants (8)
-
Alice E. Marwick -
Cornelius Puschmann -
Fred Fuchs -
Jean Burgess -
Jeff Jarvis -
Jose Marichal -
Linsey Hunter -
Prof Robyn Caplan, Ph.D.