Lisa Messeri
LightNews LightNews
banner
lmesseri.bsky.social
Lisa Messeri
@lmesseri.bsky.social
5.7K followers 1K following 790 posts
anthropologist of sci & tech. Prof @Yale. author of "Placing Outer Space" and VR book "In the Land of the Unreal". tech criticism with good vibes.
Posts Replies Media Videos
lmesseri.bsky.social
Stanford CS researchers just got a huge payday for promising AI agents that can simulate the real world. @mjcrockett.bsky.social and I wrote about these researcher's vision. Screen shotting quite a lengthy part of our paper, because we spent A LOT of time thinking about the paucity of this promise
"While most AI tries to fix humans @simile_ai is building AI that understands them. They build digital twins that capture someone’s worldview, then simulate how customers, employees or entire populations will actually respond to change. Born out of Stanford generative agent research. Now backed by $100M to turn that into a category. AI is getting smarter and Simile is making it more human. We're proud to be in their corner." A proposed solution is to build generative agents that represent specific individuals (Box 1). One such study [6] recruited a sample of ~1000 US participants nationally representative for age, gender, race, region, education, and political ideology; programmed an LLM chatbot to interview each participant for 2 h; and asked the participants to complete a battery of questionnaires and tasks. They then used the interview transcripts to prompt ~1000 LLM agents to role-play each of the human participants on the same questionnaires and tasks. Observing a high correspondence between the responses of the generative agents and their human counterparts, the researchers concluded that LLMs prompted in this way can capture the ‘idiosyncratic nature’ of real people across a range of situations [57]. Some researchers propose making generative agents even more representative by training them on their human counterparts’ ‘emails, messages and social media posts’, aswell as ‘text generated by friends, family or coworkers’ [23]. (We note this raises critical questions about informed consent; see Outstanding questions.) The logic here is that, because generative agents are built to represent a diverse sample of specific individuals, researchers could then run thousands of experiments on the generative agents and feel confident that the resultant data are faithful to the original samples. Researchers could even populate virtual worlds with generative agents, running large-scale simulations to test interventions and policies (Box 2). Nevertheless, the generative agents paradigm faces hard limits to its potential representativeness. By design, generative agents can only represent individuals who consent to sharing sensitive data with scientists, which carries substantial privacy risks [6,58]. Given these risks, people with stronger privacy concerns are less likely to consent to such studies. Members of marginalized groups in the USA, including women, gender minorities, people of color, and disabled people, have heightened privacy concerns and more negative attitudes about AI [59,60]ii–iv. These groups have historically faced disproportionate surveillance [61,62] and theft of their biometric and behavioral data for scientific research [63–65], including training machine learning models [66]. Regimes of digital surveillance spread globally [67], creating frictions where global north ideologies touch down in the global south [68]. These entrenched and repeating patterns raise cascading problems for the generative agents approach: first, members of marginalized groups are less likely to participate and, second, those who do will be less representative of their groups. Any attempt to build AI Surrogates that are truly representative of diverse populations will likely face a hard limit that marginalized people are (justifiably) less willing to entrust their data to scientists. Box 2. Generative agents and simulated worlds Researchers note that ‘many of themost interesting research questions, such as the psychology ofworld leaders, the effects of large-scale policy change, or the effects of large-scale events on the general public’ are ‘logistically infeasible’ to study in the laboratory ‘with any realistic amount of resources’ [23]. In response, generative agents populating simulated worlds are seen as promising research paths. For example, researchers could create generative agents based on the profiles of Palo Alto residents and simulate how the community would respond to different pandemic interventionsv. Much of the technical research on artificial agents acting in simulated worlds originates in fields beyond cognitive science, including computer science, sociology, economics, political science, computational social science, as well as private industry [9,112–116]. Developers of these agent architectures have lofty ambitions. They believe that this technology can ‘test interventions and theories and gain real-world insights’ [58], serving as ‘a high-fidelity platformfor policy outcome evaluation’ to enable ‘datadriven policy selection’ [115]. Given these ambitions, validating that these models can generalize to the real world is imperative [116], and some researchers caution that ‘current architectures must cover some distance before their use is reliable’ [58]. Yet, such validation faces a paradox: these models can only be validated against the ground truth of real-world data, but their appeal lies in simulating scenarios where ground truth is not available. Some researchers [22] propose to meet this challenge by identifying ‘the most proximal cases for which ground-truth data from human subjects is available’ and using those cases to validate the simulation’s predictions ‘before turning the model to a domain in which no ground truth exists’. However, there is currently ‘no consensus’ around how proximal is proximal enough [116]. Imp…
February 13, 2026 at 2:43 PM
lmesseri.bsky.social
Really enjoyed this piece by @pranavdixit.bsky.social on the demise of the Supernatural VR app. Grateful that Pranav quoted the central lesson I teach to undergrads in my "Technology and Culture" class. Highlight, tho, is the Roomba named "Mark Suckerberg."

www.businessinsider.com/zuckerberg-m...
"When I spoke to Lisa Messeri, an associate professor of anthropology at Yale and author of a 2024 book about VR, "In the Land of the Unreal," I asked her a question many Supernatural users I interviewed kept asking themselves: Is it ethical when a corporation collects years of physical and emotional data from an app that functioned as a form of self-care for thousands of people, and then pulls the rug out for a pivot? "The question is, why did we ever think that corporations were ethical in the first place?" Messeri says. For decades, the tech industry has benefited from what she calls an "endurable myth": that better technology automatically equals social progress. Messeri points to Google's original motto ,  "Don't Be Evil,"   as "the kind of grace that we gave tech companies."" Seen through that lens, the backlash against Meta over Supernatural isn't just about losing a fitness app. It's the whiplash of thousands of people realizing that something they treated as a routine, a community, and even a mental health lifeline, was never anything more than a plaything of billionaires. "Part of the anger that you're seeing "is coming from this feeling of betrayal that comes with an enduring belief in that myth," Messeri says.
February 3, 2026 at 1:55 PM
lmesseri.bsky.social
I watched, uh, a lot of these keynotes. For science. But these watchings and re-watchings made me realize the increasingly fictive qualities of what used to be straightforward demos 5/
screen shot from metaverse keynote, with robots, floating avatars and facetime chat screenshot from Apple's spatial computing keynote. woman standing in front of a vignette illustrating at home use of tech.
January 29, 2026 at 6:40 PM
lmesseri.bsky.social
oops, w/ alt text
Visions of Computing Futures: Meta, Apple, and the Technological Fictions of Digitally Real Worlds Abstract: The vision of seamlessly integrating virtual and physical environments – variously called spatial computing and the metaverse – has been around for decades. Recently, Big Tech companies, including Apple and Meta, have prioritized ushering in this computing paradigm. Despite hardware and interface inadequacies, how do these companies make such visions feel not only inevitable but also readily achievable? This paper analyzes performances of corporate vision-making as technological fictions, drawing attention to the genre and aesthetics that make a future seem like the future. By interweaving narratives illustrating the sociality of metaverse futures with presentations of technical capabilities, corporate visions purposely blur fiction and fact such that the former is mistaken for the latter. As scholars consider how to refuse Big Tech visions, excavating how inevitabilities are mere possibilities offers one critical strategy.
December 5, 2025 at 4:44 PM
lmesseri.bsky.social
In a stroke of good timing, my article excoriating the metaverse (and taking shots at spatial computing) was just accepted at Games & Culture. How do these companies make implausible and undesirable futures seem inevitable? abstract posted here!
December 5, 2025 at 4:42 PM
lmesseri.bsky.social
#AAA2025 has been its usual whirlwind of fun and overwhelm. But I can’t underscore how much it has meant for ‘in the land of the unreal’ to be recognized as the winner of the Bateson book prize, and to celebrate the news w my ever growing anthro fam. @dukepress.bsky.social @culanth.bsky.social
Sign announcing ‘ln the land of the unreal’ as Bateson prize winner
November 22, 2025 at 8:53 PM
lmesseri.bsky.social
absolutely. which is why this was the definition for AI that @mjcrockett.bsky.social and I included in our recent TiCS article
domain of multidisciplinary academic inquiry that dates back to the mid-20th century. It has more recently become a successful marketing term deployed by large tech companies to describe a host of automated and semiautomated processes. Given that the purposely underdefined and hype-fueled meaning of AI has taken over more academic definitions, we limit our use of the term to capturing social trends and impulses rather than for referring to a scientific body of work.
November 11, 2025 at 2:00 PM
lmesseri.bsky.social
With Valerie Olson, we wrote "Spaced Out: Bringing Outer Space into Anthropological Conversations." We observe: "When anthropologists include outer spaces within anthropological space, they surface hidden binaries that shape anthropological conceptualizations of 'space' as a general category."
cover of the book "Otherwhere ethnography: an introduction to outer space studies." edited by Istvan Praet and Perig Pitrou
November 6, 2025 at 2:25 PM
lmesseri.bsky.social
I hope you enjoy the paper! There are fun easter eggs hidden throughout, particularly the definition of AI that somehow made it past peer and editorial review and into the glossary. 13/
Artificial Intelligence (AI) domain of multidisciplinary academic inquiry that dates back to the mid-20th century. It has more recently become a successful marketing term deployed by large tech companies to describe a host of automated and semiautomated processes. Given that the purposely underdefined and hype-fueled meaning of AI has taken over more academic definitions, we limit our use of the term to capturing social trends and impulses rather than for referring to a scientific body of work.
October 21, 2025 at 3:47 PM
lmesseri.bsky.social
The DEAD critique comes from observing the historical arc of experimental work, noting that the rise of computers, then MTurk, now LLMs has increasingly constrained the kinds of cognitive tasks that research participants perform. 7/
a funnel showing As investigations of cognition transition from the world to laboratory studies (blue), to computerized studies (red), to online studies (orange), to AI Surrogate studies (yellow), the varieties of cognition that are possible to measure become narrowed to Decontextualized, Engineered, Anonymized, Disembodied (DEAD) cognition.
October 21, 2025 at 3:47 PM
lmesseri.bsky.social
We write, the “conditions upon which AI tools are built inherently exclude the same people who are already underrepresented in CogSci… AI Surrogates perpetuate an illusion of generalizability where researchers believe these models represent a broader swathe of humanity than they actually do.” 5/
a figure from the paper, illustrating illusions of generalizability. in the top, well-recognized illusion of generalizability arises when researchers mistake Western, Educated, Industrialized, Rich, Democratic (WEIRD) convenience samples as more representative of all humans than they actually are. The illusion is obscured when we describe convenience samples as ‘people’ or ‘humans’ in the papers we write. in the bottom, nother illusion of generalizability arises when researchers mistake DEAD cognitive processes as being more representative of all cognition than they actually are. This illusion is obscured when we write papers about abstracted cognitive processes rather than about the more specific forms of cognition that DEAD tasks measure.
October 21, 2025 at 3:47 PM
lmesseri.bsky.social
In our nature paper, we discussed four visions of AI in science: the Oracle, the Surrogate, the Quant, and the Arbiter. This is a deep dive into the surrogate 🤖 - the promise that LLMs can serve as human research participants 2/
The abstract reads "Recent advances in artificial intelligence (AI) have generated enthusiasm for using AI simulations of human research participants to generate new knowledge about human cognition and behavior. This vision of ‘AI Surrogates’ promises to enhance research in cognitive science by addressing longstanding challenges to the generalizability of human subjects research. AI Surrogates are envisioned as expanding the diversity of populations and contexts that we can feasibly study with the tools of cognitive science. Here, we caution that investing in AI Surrogates risks entrenching research practices that narrow the scope of cognitive science research, perpetuating ‘illusions of generalizability’ where we believe our findings are more generalizable than they actually are. Taking the vision of AI Surrogates seriously helps illuminate a path toward a more inclusive cognitive science."
October 21, 2025 at 3:47 PM
lmesseri.bsky.social
No kings
October 18, 2025 at 4:03 PM
lmesseri.bsky.social
The president of MIT has declined to sign "the compact". Full letter, as shared with the MIT community:
Dear Madam Secretary, I write in response to your letter of October 1, inviting MIT to review a “Compact for Academic Excellence in Higher Education.” I acknowledge the vital importance of these matters. I appreciated the chance to meet with you earlier this year to discuss the priorities we share for American higher education. As we discussed, the Institute’s mission of service to the nation directs us to advance knowledge, educate students and bring knowledge to bear on the world’s great challenges. We do that in line with a clear set of values, with excellence above all. Some practical examples: • MIT prides itself on rewarding merit. Students, faculty and staff succeed here based on the strength of their talent, ideas and hard work. For instance, the Institute was the first to reinstate the SAT/ACT requirement after the pandemic. And MIT has never had legacy preferences in admissions. • MIT opens its doors to the most talented students regardless of their family’s finances. Admissions are need-blind. Incoming undergraduates whose families earn less than $200,000 a year pay no tuition. Nearly 88% of our last graduating class left MIT with no debt for their education. We make a wealth of free courses and low-cost certificates available to any American with an internet connection. Of the undergraduate degrees we award, 94% are in STEM fields. And in service to the nation, we cap enrollment of international undergraduates at roughly 10%. • We value free expression, as clearly described in the MIT Statement on Freedom of Expression and Academic Freedom. We must hear facts and opinions we don’t like – and engage respectfully with those with whom we disagree. These values and other MIT practices meet or exceed many standards outlined in the document you sent. We freely choose these values because they’re right, and we live by them because they support our mission – work of immense value to the prosperity, competitiveness, health and security of the United States. And of course, MIT abides by the law. The document also includes principles with which we disagree, including those that would restrict freedom of expression and our independence as an institution. And fundamentally, the premise of the document is inconsistent with our core belief that scientific funding should be based on scientific merit alone. In our view, America’s leadership in science and innovation depends on independent thinking and open competition for excellence. In that free marketplace of ideas, the people of MIT gladly compete with the very best, without preferences. Therefore, with respect, we cannot support the proposed approach to addressing the issues facing higher education. As you know, MIT’s record of service to the nation is long and enduring. Eight decades ago, MIT leaders helped invent a scientific partnership between America’s research universities and the U.S. government that has delivered extraordinary benefits for the American people. We continue to believe in the power of this partnership to serve the nation. Sincerely, Sally Kornbluth cc Ms. May Mailman Mr. Vincent Haley
October 10, 2025 at 2:42 PM
lmesseri.bsky.social
i'm much madder at Big Tech than i was when i first submitted this article. I offer my tracked changes revision as evidence of entering the no holds bar stage of my career.
And yet, both of these corporate futures have far more in common than in distinction. Especially as spatial computing has, like the metaverse, failed to garner the enthusiasm corporate overloads had anticipated, it is important to understand these terms in the context of the larger computing visions they seek to proffer. This vision wants users to be permanently and inescapably immersed in a world of digital overlays, such that the corporate product, in whatever technical form (headset, glasses), is the mediating screen through which we always and only experience the physical world. This is a vision we ought to refuse, and this article offers an admittedly modest strategy for refusal. I do so by offering a reading of the technological fictions that convey the spatialities and socialities of spatial computing
October 3, 2025 at 9:46 PM
lmesseri.bsky.social
Turns out, Mapquest still labels the body of water to the west of Florida The Gulf of Mexico. Welcome to the resistance, Mapquest.
screen shot of a map, showing that the "gulf of Mexico" has not been relabled.
September 15, 2025 at 1:08 PM
lmesseri.bsky.social
i check in with ChatGPT sometimes to see where it's at as a research tool. I asked it some questions about a current project. i was at first dishearteningly impressed. It found the main stuff i found (in minutes as opposed to weeks). But, within 6 prompts it was hallucinating. I have thoughts...
a screenshot of a chatGPT interface where it admits to making up a source.
July 16, 2025 at 7:21 PM
lmesseri.bsky.social
Regarding the "AI for Good" conference presently going on, here's a short passage from "In the Land of the Unreal" (pg 9). This idea is drawing directly from Magalhães and @couldrynick.bsky.social's article “Giving by Taking Away.”
"Because “the good” is not a universally agreed upon concept but rather articulated through “regimes of knowledge,” “tech for good” projects that emerge from Big Tech companies shape definitions of what it means to do “good” work in ways that align with their mission, namely capital accumulation"
July 10, 2025 at 2:47 PM
lmesseri.bsky.social
Disgusted by this Zoom UI: when i mouse over the link icon to copy a zoom link, it is replaced by an "AI companion" link.

Google was doing this with gemini for awhile as well. This is just nasty UI design, tricking people to click on a feature that they (I) don't want.
screen shot of a zoom interface with a link icon screenshot showing how the link icon is bumped over to make room for an AI companion
July 9, 2025 at 9:45 PM
lmesseri.bsky.social
ah, so that's how you quit the gym.
screen shot with the following text "You may also cancel this contract if: You relocate your residence further than twenty-five (25) miles from any health club operated by the seller or from any other substantially similar health club which would accept the obligation of the seller; You die; or"
July 8, 2025 at 7:07 PM
lmesseri.bsky.social
Threaten us with a good time, why don’t you
Trump threatens to cut musk’s government contracts as ties unravel
June 5, 2025 at 7:11 PM
lmesseri.bsky.social
"Teaching with ChatGPT," an essay i wrote in 2023 reflecting on the first post-ChatGPT semester, holds up unfortunately well. www.tandfonline.com/doi/abs/10.1...
My third concern, however, makes me fear that there might be little we can do to curtail students’ instincts to incorporate AI into their workflow. Just as I was able to take a mediocre lesson plan and make it something fun and exciting, there are no doubt students who have performed similar alchemy and bragged to their classmates about receiving a top mark and doing less work. Given how accessible (and free) these apps are, students might come to feel that it is necessary to use AI both to keep up with the pack and, should faculty get used to AI-inflected products, adapt to the new normative style of the pack. Educators have been vocal in their criticism of GPT, and we can see how easily it will be used to replace, rather than assist, the work we desire our students to complete. Further, given its penchant for returning wrong information, I fear that our students will be conducting research in a digital environment where truths and realities are not a prerequisite for authoritative responses. While using AI in college might make our students better prepared to enter an AI-infused workforce, the trade-off will be that they might lose the ability to critically interrogate whether this is a workforce they even desire to enter.
May 8, 2025 at 12:58 PM
lmesseri.bsky.social
Final projects for ‘Anthropology of Outer Space’ are on view until 4 at Yale Peabody Museum. My students built a world where there are humans on Mars. These are their things.
Sign for ‘Mars unearthed’ exhibit Plastic cups painted to look like different planets as part of the sun shrine CDs of Mars music Magma supplements for all your Martian needs
April 29, 2025 at 5:40 PM
lmesseri.bsky.social
ffs indeed. this is the second time this ludicrous higher has pushed a news cycle. here's what i said to business insider last time. holds true....
April 24, 2025 at 4:49 PM
lmesseri.bsky.social
scream. scream scream scream.

Perhaps most offensive, "Ask ChatGPT to do detailed online research, saving you hours of time."

Saving you time.... to do what? it's finals week? this is what students are pedagogically supposed to be doing during this period.
Use ChatGPT Plus to help you through finals: Upload dozens of docs, PDFs, and images and ask ChatGPT to summarize or quiz you Ask ChatGPT to do detailed online research, saving you hours of time Brainstorm writing ideas and get edits and writing tips, too Practice speaking foreign languages – or prep for interviews – with longer real-time conversations Create custom images for presentations and projects
April 15, 2025 at 5:19 PM