Jean-François Bonnefon
@jfbonnefon.bsky.social
Behavioral Science of AI @ Toulouse School of Economics @tse-fr.eu Director of @iast.fr Chair of Moral AI @ Artificial and Natural Intelligence Toulouse Institute https://jfbonnefon.github.io/ I have a successful life with bipolar disorder
French Theory, the collectible card game edition. I got a Kristeva and a Lacan and they have ok powers, but good luck getting a Foucault or a Bourdieu
November 9, 2025 at 3:11 PM
French Theory, the collectible card game edition. I got a Kristeva and a Lacan and they have ok powers, but good luck getting a Foucault or a Bourdieu
It can be hard for multidisciplinary teams to agree on the title of their manuscripts
November 8, 2025 at 12:33 PM
It can be hard for multidisciplinary teams to agree on the title of their manuscripts
Please select six keywords for your manuscript
November 8, 2025 at 11:53 AM
Please select six keywords for your manuscript
As seen in the release notes of ChatGPT Atlas. Cool cool cool
October 22, 2025 at 10:07 AM
As seen in the release notes of ChatGPT Atlas. Cool cool cool
A professional fact-checker vetted every modified headline for accuracy. Positive engagement with modified climate headlines increased by ~10pp among skeptics, and the effect was strongest among the most skeptics +
October 13, 2025 at 9:13 AM
A professional fact-checker vetted every modified headline for accuracy. Positive engagement with modified climate headlines increased by ~10pp among skeptics, and the effect was strongest among the most skeptics +
published today in @natclimate.nature.com with @bencebago.bsky.social & @philmuller.bsky.social
We used Llama-3 70B to subtly rewrite climate headlines and make them palatable for skeptics. 2,000 US participants upvoted, downvoted, bookmarked headlines in a reddit-styled experiment. +
We used Llama-3 70B to subtly rewrite climate headlines and make them palatable for skeptics. 2,000 US participants upvoted, downvoted, bookmarked headlines in a reddit-styled experiment. +
October 13, 2025 at 9:13 AM
published today in @natclimate.nature.com with @bencebago.bsky.social & @philmuller.bsky.social
We used Llama-3 70B to subtly rewrite climate headlines and make them palatable for skeptics. 2,000 US participants upvoted, downvoted, bookmarked headlines in a reddit-styled experiment. +
We used Llama-3 70B to subtly rewrite climate headlines and make them palatable for skeptics. 2,000 US participants upvoted, downvoted, bookmarked headlines in a reddit-styled experiment. +
Taylor Swift singing about peer review
October 7, 2025 at 10:20 AM
Taylor Swift singing about peer review
what a playful cover featuring examples of what participants told the AI to do in our ‘delegation to AI can increase dishonest behavior‘ article
www.nature.com/articles/s41...
www.nature.com/articles/s41...
October 2, 2025 at 6:53 AM
what a playful cover featuring examples of what participants told the AI to do in our ‘delegation to AI can increase dishonest behavior‘ article
www.nature.com/articles/s41...
www.nature.com/articles/s41...
ok ok so Glinda uses ChatGPT but Elphaba doesn't?
August 23, 2025 at 12:54 PM
ok ok so Glinda uses ChatGPT but Elphaba doesn't?
Do I get a new tote bag for each paper I review? If not, what is my incentive to do a second review?
August 14, 2025 at 4:46 PM
Do I get a new tote bag for each paper I review? If not, what is my incentive to do a second review?
Our adapted futures cone: sweet spot = plausible, near-term, moderate-impact tech; too remote, uncertain, or disruptive tanks validity, speaking of which +
August 6, 2025 at 3:26 PM
Our adapted futures cone: sweet spot = plausible, near-term, moderate-impact tech; too remote, uncertain, or disruptive tanks validity, speaking of which +
Spectrum of simulation methods: simple text vignettes → interactive apps → AR/VR testbeds → physical simulations. Pick the fidelity your question really needs +
August 6, 2025 at 3:26 PM
Spectrum of simulation methods: simple text vignettes → interactive apps → AR/VR testbeds → physical simulations. Pick the fidelity your question really needs +
Motivation is the Collingridge dilemma: easiest to steer tech before it launches, but hardest to predict its impact at that point. sci-fi-sci tackles that gap head-on +
![Title: Examples of policy debates about nascent or speculative technologies.
1. Connected and autonomous vehicles (CAVs)
Autonomous driving technology is still being tested, and fully autonomous cars are confined to specific cities. Yet, a debate ensued early on about the ethical implications of intelligent machines making life-and-death decisions in unavoidable accident scenarios [66]. These debates have led to significant mobilization by Civil Society [67], and specific policy recommendations [26], some of which have informed legislation [68].
2. Social credit systems
Governments do not yet have the capacity to deploy AI systems that would monitor every behavior of all their citizens in real time, in order to calculate and release social scores; but the behavioral, social, and political impact of these hypothetical systems is already the object of much speculation [69, 70, 71, 72], to the point where the European Union is considering a preemptive ban of this technology for its member governments [73].
3. Embryo screening
Fertility clinics are only in the very early stages of providing parents with the option to choose or reject specific genetic traits in their offspring [74]. However, bioethicists are already considering the ethical trade-offs, social dilemmas, and collective action problems that may arise if parents gain the ability to selectively determine cognitive, moral, physical, and immune traits [75, 76, 77].
4. Ectogenesis
Medical research has not yet achieved ectogenesis (i.e., the artificial gestation of human fetuses), but the field of speculative bioethics is already debating whether this potential technology might end gender-based oppression, or exacerbate other inequalities, and to which extent it could disrupt social and legal norms around motherhood and employment rights [78, 79, 80, 81].](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:pm6o2vqrg2yd2jm7qwcigfjl/bafkreihpa25tmwgafocbgibwiznodwdarxx6sgcgwkgs2mbujiqsq3jwju@jpeg)
August 6, 2025 at 3:26 PM
Motivation is the Collingridge dilemma: easiest to steer tech before it launches, but hardest to predict its impact at that point. sci-fi-sci tackles that gap head-on +
Is it just me who thinks of Nature the journal, whenever people say something inspiring about nature, the thing outside
July 26, 2025 at 12:36 PM
Is it just me who thinks of Nature the journal, whenever people say something inspiring about nature, the thing outside
I have no idea what this is about but I’ve never seen such a badass title
July 2, 2025 at 7:27 AM
I have no idea what this is about but I’ve never seen such a badass title
my wife makes the most stylish scarecrows
May 29, 2025 at 1:41 PM
my wife makes the most stylish scarecrows
There are many many such papers, I'm sure @mariejuanchich.bsky.social know them all and can point you to a nice review; but maybe the plot you are thinking of is this one, not from a paper but created for Reddit, it became quite viral
May 15, 2025 at 11:17 AM
There are many many such papers, I'm sure @mariejuanchich.bsky.social know them all and can point you to a nice review; but maybe the plot you are thinking of is this one, not from a paper but created for Reddit, it became quite viral
The quote marks
April 2, 2025 at 12:40 PM
The quote marks
This bathroom welcomes all genders, disabilities, and Olympic swimmers
March 23, 2025 at 11:08 AM
This bathroom welcomes all genders, disabilities, and Olympic swimmers
There is some wild cyberpunk stuff in the EU document explaining prohibited AI uses, like video games that can hack your brain in order to guess your PIN
February 12, 2025 at 10:53 AM
There is some wild cyberpunk stuff in the EU document explaining prohibited AI uses, like video games that can hack your brain in order to guess your PIN
Un peu brutal comme intro quand même
January 28, 2025 at 1:49 PM
Un peu brutal comme intro quand même
