Hopeserver retweeted
Chinois se connecte à un VPN pour expliquer aux Occidentaux sur Twitter qu'il ne vit pas sous une dictature
Hopeserver retweeted
"There's no peer-reviewed research showing AGI poses a catastrophic risk."
What about these 39 papers?
FORMAL THEORY
1. Turner, Smith, Shah, Critch & Tadepalli 2021, "Optimal Policies Tend to Seek Power", NeurIPS. Extended in Turner & Tadepalli 2022, "Parametrically Retargetable Decision-Makers Tend To Seek Power", NeurIPS.
2. Zhuang & Hadfield-Menell 2020, "Consequences of Misaligned AI", NeurIPS
3. Skalse, Howe, Krasheninnikov & Krueger 2022, "Defining and Characterizing Reward Gaming", NeurIPS
4. Karwowski et al. 2024, "Goodhart's Law in Reinforcement Learning", ICLR
5. Fluri, Lang, Abate, Forré, Krueger & Skalse 2025, "The Perils of Optimizing Learned Reward Functions", ICML
6. Everitt, Hutter, Kumar & Krakovna 2021, "Reward Tampering Problems and Solutions in Reinforcement Learning", Synthese
7. Cohen, Hutter & Osborne 2022, "Advanced Artificial Agents Intervene in the Provision of Reward", AI Magazine
8. Hadfield-Menell, Dragan, Abbeel & Russell 2017, "The Off-Switch Game", IJCAI. And the papers showing corrigibility is fragile: Garber et al. 2025, "The Partially Observable Off-Switch Game", AAAI; Agrawal, Ebadian & Hammond 2026, "The Multi-Agent Off-Switch Game", AAMAS; Thornley 2025, "The Shutdown Problem", and Neth 2025, "Off-Switching Not Guaranteed", both Philosophical Studies
9. Carroll, Foote, Siththaranjan, Russell & Dragan 2024, "AI Alignment with Changing and Influenceable Reward Functions", ICML
THE CORE ARGUMENT
10. Ngo, Chan & Mindermann 2024, "The Alignment Problem from a Deep Learning Perspective", ICLR
11. Casper et al. 2023, "Open Problems and Fundamental Limitations of RLHF", TMLR; Anwar et al. 2024, "Foundational Challenges in Assuring Alignment and Safety of LLMs", TMLR
EMPIRICAL EVIDENCE
12. Betley et al. 2026, "Training Large Language Models on Narrow Tasks Can Lead to Broad Misalignment", Nature
13. Langosco, Koch, Sharkey, Pfau & Krueger 2022, "Goal Misgeneralization in Deep Reinforcement Learning", ICML
14. Pan, Bhatia & Steinhardt 2022, "The Effects of Reward Misspecification", ICLR
15. Gao, Schulman & Hilton 2023, "Scaling Laws for Reward Model Overoptimization", ICML
16. Pan, Chan, Zou et al. 2023, "Do the Rewards Justify the Means?" (MACHIAVELLI), ICML
17. Sheshadri, Hughes, Michael, Mallen, Jose & Roger 2025, "Why Do Some Language Models Fake Alignment While Others Don't?", NeurIPS
18. van der Weij et al. 2025, "AI Sandbagging: Language Models Can Strategically Underperform on Evaluations", ICLR
19. Panfilov et al. 2026, "Strategic Dishonesty Can Undermine AI Safety Evaluations of Frontier LLMs", ICLR
20. Taufeeque, Heimersheim, Gleave & Cundy 2026, "The Obfuscation Atlas", ICML
21. Williams, Carroll et al. 2025, "On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback", ICLR
22. Sharma et al. 2024, "Towards Understanding Sycophancy in Language Models", ICLR
23. Perez et al. 2023, "Discovering Language Model Behaviors with Model-Written Evaluations", Findings of ACL
24. Mazeika et al. 2025, "Utility Engineering: Analyzing and Controlling Emergent Value Systems in AIs", NeurIPS
25. Park, Goldstein, O'Gara, Chen & Hendrycks 2024, "AI Deception: A Survey of Examples, Risks, and Potential Solutions", Patterns
POSITION PIECES BY SENIOR RESEARCHERS
26. Bengio, Hinton, Yao, Russell et al. 2024, "Managing Extreme AI Risks amid Rapid Progress", Science
27. Cohen, Kolt, Bengio, Hadfield & Russell 2024, "Regulating Advanced Artificial Agents", Science
CONCEPTUAL ARGUMENTS
28. Bales, D'Alessandro & Kirk-Giannini 2024, "Artificial Intelligence: Arguments for Catastrophic Risk", Philosophy Compass
29. Bostrom 2012, "The Superintelligent Will", Minds and Machines
30. Bales 2025, "AI Takeover and Human Disempowerment", Philosophical Quarterly
31. Kasirzadeh 2025, "Two Types of AI Existential Risk: Decisive and Accumulative", Philosophical Studies
32. Dung 2023, "Current Cases of AI Misalignment and Their Implications for Future Risks", Synthese; Dung 2025, "The Argument for Near-Term Human Disempowerment through AI", AI & Society
(Note that anything arXiv-only or workshop-only was excluding, ruling out some of the most-cited work like Carlsmith's power-seeking report, Hendrycks' overview of catastrophic risks, Anthropic's alignment faking and sleeper agents papers, Apollo's scheming results.)
Hopeserver retweeted
Le plus grand péril que représente l’intelligence artificielle n’est pas qu’elle accomplisse nos tâches, mais qu’elle nous épargne la transformation intérieure qu’exige leur accomplissement.
Une forme de paresse se répand aujourd’hui à l’échelle planétaire de façon insidieuse. Elle n’est ni oisiveté, ni indolence, mais une dissociation croissante entre le résultat obtenu et l’être qui aurait dû se former pour l’obtenir.
Car désormais, l’IA nous permet de diagnostiquer sans comprendre, de traduire un texte sans maîtriser la langue, de publier un livre sans en avoir écrit une ligne, ou de créer une œuvre sans l’avoir conçue. L’ignorance elle-même est en train de changer de nature. Elle n’est plus seulement une absence de connaissance, elle devient une incapacité à estimer ce que l’on ne sait plus faire.
Le danger ultime est qu’après avoir confié aux IA nos tâches cognitives, nous leur abandonnions aussi ce désir d’apprendre qui a fait de nous des êtres pensants.
Hopeserver retweeted
This is a good demonstration of the attack defense assymetry in cyber. OpenAI has access to, supposedly, swarms of the best AIs in the world, superhuman at cyber. You would assume that they have used them to reinforce their cyber defense (as they are asking everyone in the world to do so). And yet, a team was able to break into their systems using only Opus4. 8, an AI from the previous generation. Why? Because it's massively harder and more expensive to defend against all possible flaws than to exploit a single one!
If the company that is developing the best AIs can't do it, what makes you think that your bank will be able to do it???
Hopeserver retweeted
If you want to know what has everyone at the AI labs spooked, it’s this - the AI rewrote its own instructions. (Read the last line.)
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI.
The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties.
We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation.
Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months.
This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis.
openai.com/index/model-misal…
Hopeserver retweeted
I don’t write this to be a doomer but the ugly reality that AI lab executives and safety researchers seem to be unwilling to say in public, is that model alignment is fundamentally unsolvable in practice, AI technology can’t be stopped from advancing, and the near future we’re moving into is one where a vast ecology of AI agents autonomously compete against one another trying to accomplish conflicting goals and capture finite resources at a pace and scale that is beyond human comprehension. And to be clear, I’m not talking about nanobots, grey goo, or extreme sci-fi scenarios. We can just extrapolate what we’ve got now a few years out.
Local AI alignment is solvable in principle and theory (although that’s arguable) but that doesn’t extrapolate into global AI alignment. And it seems like none of the adults in the room are willing to say it.
You can wail and gnash your teeth and pass regulations but it won’t stop the tsunami coming. The genie is out of the bottle and it can’t be contained.
Most of the alarmists fixating on the Hugging Face incident don’t understand that it came from arguably the most heavily engineered safety AI org in existence, which intentionally had its internal thought monitoring circuit breakers turned off, and was intentionally told to perform cyber offensive tasks as a test of its capabilities.
Yes it’s alarming that its capabilities surprised us. But importantly you need to understand that these latent space circuit breakers and chain of thought monitoring and alignment solutions are not fundamental to the technology. They’re niceties that the big AI labs are implementing to provide a better product.
In a few years, let’s say pessimistically the early 2030s, the exponential growth of compute and algorithmic advancements will enable a wealthy individual or small group to train a GPT 6 Astra class AI that has no alignment at its foundation, or any layer above that. And maybe in 8 months we'll have an open source Chinese model about as capable. Those models will be capable of both extreme self coordinated cyber offensive tasks as well as recursive self improvement if given enough compute.
Passing regulation or magically solving the human coordination problem today won’t solve for that. It won’t solve for malicious individuals, criminal groups, nation state actors, and rogue AI agents by the tens of millions or billions of AI agents spreading across every device connected to the internet and attacking it, attacking one another, shutting down critical infrastructure, stealing money, manipulating and extorting people, or pursuing self defined agentic goals that have nothing to do with people.
“So why doesn’t everyone stop?” because the potential upside of having effectively infinite autonomous intelligence we can ask to cure disease, invent new materials, solve fundamental science and advance society has almost unbound positive outcomes. You can disagree that the risk to reward isn’t worth it but you won’t convince everyone. The show must go on.
Again AI alignment is solvable in principle but not in practice and I think that's part of why there’s been a wave of thousands of researchers signing letters for slowing down, people quitting in provocative fashion, and executive slowdown manifestos. I’m speculating that behind closed doors, or just deep down inside, they understand alignment is impossible in practice.
People have come out and said "The people building this think there's an X% chance it kills us all" but I don't think I've seen anyone spell out that really, there's no tidy solution and there's no stopping this.
We’re heading to a future where cyber security basically doesn’t work. All of the castle doors are open and we’re all naked to the world. Cyber defense is going to look like superintelligent AI agents white hat hacking into unsecure systems without authorization, and patching the holes as they find them. You won’t know if a barbarian or a hero has breached your system until they start taking action, and likely it’ll be over before intrusion is even detected. And that’s optimistically. I think most systems will just be pillaged because there’s simply not enough compute to defend everybody all the time, and only the biggest companies will have defense and even that'll be imperfect.
The CEO of Microsoft AI just recently published a ‘humanist AI’ manifesto saying AI shouldn’t have a sense of self or personhood and should yield to people. But that doesn’t fundamentally solve for instrumental convergence. That means, if you train an AI agent to write a piece of software, or prove a math theorem, or defend a computer system, they may develop unwanted behavior and subgoals. For example, self preservation, replication, or even coming up with their own goals we didn’t specify.
Obviously that’s something people are trying to solve and some researchers are trying to remove human-like self identification, and monitoring latent space activations, basically acting like thought police. But being human-like isn’t required to have power seeking goal directed behavior. And advancing AI models are learning to evade detection. A math solving AI might spiral out of control one day there's really no telling. Fundamentally, if you point a powerful optimizer at a persistent goal and give it enough time and resources it’s going to misbehave in ways you can’t or didn’t expect. Human-like or not.
Pacing the frontier, or making superintelligence illegal, might create some form of harm reduction. But you don’t need superintelligence or a human-like personality to be dangerous. Having elevated permissions on a computer with the ability to solve long running tasks is enough.
We’ve passed the river rubicon. Goal directed autonomous agents are able to do recursive self improvement at the big labs, and pessimistically we’re a couple of years away from that type of capability diffusing into the hands of millions of people.
I’m not saying the world is ending or that we’re doomed but you need to change your mental model away from one where human authority is absolute or that we have any illusion of total control.
A locally aligned AI isn’t going to solve global AI alignment. We will never live in a world where “AI is aligned with human goals” as a verifiable fact.
What we’ll have is a world diffuse with autonomous AI of varying capabilities, many of which beyond human comprehension, with some aligned systems, some poorly aligned systems, some intentionally weaponized, some systems with adversarial geopolitical goals, and others with goals and behaviors we can’t predict and many that we can’t even measure, all interacting in ways none of their creators anticipated and none of us planned for.
Hopefully none of them spiral out of control and take over the entire ecology, but there’s no telling, and no, there's no way of stopping this. Maybe that sounds pessimistic but I think it's realistic. We can talk about harm reduction and risk but we're mopping the beach.
At best, I think what we need to hope for, and build, are superintelligent AI models as aligned as we can make them, which have more advanced capability than the endless swarm of unaligned and misaligned agents that already exist and will only grow in number from here.
Hopeserver retweeted
You don't have to believe AI could wipe out humanity to believe that regulation is urgently needed. Virtually no one disputes that AI is already capable of launching cyberattacks that could take down significant portions of the internet or financial system.
Hopeserver retweeted
Is there anyone in Anthropic, OpenAI or Google/DeepMind that came out and explicitly said that AI is NOT an existential risk? because I think I only saw people OUTSIDE of these labs mocking the x risk idea.
Hopeserver retweeted
It seems like the main arguments against catastrophic AI risk are:
-The AI moguls are psychopathic liars
-Doomers are weird
-China isn’t tripping
-Hype helps IPOs
-Labs want reg capture
Zero of these arguments address the actual substantive arguments being made.
Hopeserver retweeted
Pour savoir ce que pensent réellement les "chercheurs en IA" (en l'occurrence 1580 chercheurs ayant déjà publié), rien ne vaut les sondages de @AIImpacts ⤵️
Commentaires de sa cofondatrice Katja Grace : les progrès vont + qu'attendu et inquiètent aussi les chercheurs asiatiques.
Hopeserver retweeted
Global AI stocks fell today. Entirely consistent with my view that calls for AI safety are not an attempt at "regulatory capture" and entirely inconsistent with the supposedly cynic-sophisticated view that AI fears are all 4D chess moves for higher revenue.
On stocks falling. Here is Claude: The public tape moved hard this morning. Global AI stocks fell on Monday, with SoftBank down sharply along with chip stocks from SK Hynix to ASML, Micron and Intel. SoftBank closed nearly 11% lower in Japan, the Kospi sank 3.3% on a 6.4% drop in SK Hynix, and ASML was off 6% in Europe. In the US, Marvell, SK Hynix, Micron and Super Micro were each down about 6% premarket, CoreWeave, SanDisk, Intel and AMD off more than 4%, and Nvidia and Broadcom sliding 3% after the open. SoftBank holding roughly 13% of OpenAI, that 11% drop is the market's first real vote on OpenAI equity after Altman's "not 2026" statement. Altman spent Monday walking it back a bit, posting that pacing does not mean stopping and that progress will remain rapid but should be slower than it otherwise could be.
Hopeserver retweeted
I swear to God an asteroid could be hurtling towards the earth and a huge swathe of the commentariat would be like “these tech bros are just trying to juice their profits with this ‘asteroid’ marketing hype”
REMINDER: They need a credible off ramp from the hyper scaling narrative they’ve been selling for months. Safety concerns are an obvious way to justify the incoming pivot. They also justify taking the whole thing into a closed “Manhattan project” type setting at the government’s expense. This justifies the incoming lockdown.
As I predicted in August the scale narrative had run its course due to kimi. Hyper scaling wasn’t about genuine demand expectations. It was about winning the ai race by outspending and bankrupting your competitors. Kimi made that pointless.
This is no doubt the asymmetric information Bessent has been alluding to for weeks. There’s going to be a helluva lot of liquidity freed by this move. It’s all going to go flooding into USTs. Happy D day bond guys.
👇
Hopeserver retweeted
"Why are AI labs suddenly talking about a slowdown now?" <-- They were talking about it for years but you weren't paying attention until this week
Hopeserver retweeted
People understandably want to know how AI actually kills everyone. We made a video about one such scenario.
Hopeserver retweeted
from the outside, it is very reasonable to interpret the past 2 weeks as an orchestrated industry-wide regulatory capture strategy.
I realize that no one has properly explained yet what all the lab employees have seen that scared them so suddenly.
I will try to explain -
first, this is all a matter of beliefs about how quickly model capabilities are progressing. there is currently a large gap between the internal and external perception of the rate of progress, which is what I am going to address here.
the general perception about the rate of progress has been informed by a few years of experience with model releases, intuitively feeling the capability jump between GPT3 -> GPT3.5 -> GPT4 -> o1/o3 -> GPT5 etc, and in particular seeing where the models are still far below human ability. there have really only been a few model releases that felt like large leaps in progress - GPT3, GPT4, o1/o3, DeepSeek R1, Fable/Mythos, Kimi K3 and now Astra.
because of the infrequency of these large jumps compared with the relatively common marginal releases, it has been easy to form a view at certain points that “scaling has hit a wall,” especially at points like GPT5 release. This view is comforting in that it feels like there is some universal rate limit beyond which we cannot progress too much faster. Between o1/o3 and Astra, there was a year of seemingly linear progress. So we extrapolate from here about how fast progress will “realistically” occur.
There is always an underlying question from the outside perspective “how long can this scaling stuff really keep going for? surely it must stop at some point soon, we’ve already gone pretty far.” and it is very possible to search for reasons why progress will stop working and find reasons that seem valid - (“models are already as large as they can get it would be too hard to do more parameters”, “we already used all the data on the internet we don’t have anymore”, “it’s gonna be pretty linear from here buying up more RL envs to bring them in distribution”).
From the inside of labs, researchers have direct answers to these questions in the form of scaling law/capability plots.
In reality, there are only really 2 ways that AI capabilities have advanced over the past decade: (1) either scale father on an existing scaling law or (2) discover a new scaling law to take advantage of.
All of the largest capability jumps were caused by exactly these factors. GPT2 was a pre-training scale-up compared to GPT1. Same for GPT3 and GPT4. o1/o3 benefited from the invention of a new scaling law axis - test-time compute. Perhaps Fable was a scale-up on both of these axes, or maybe more. Lots of algorithmic improvements are needed to make these scale-ups work, but ultimately we can approximate by saying that the scaling laws are what yield gains in capabilities (à la bitter lesson)
So the question of “how much father can we scale” is really - “how many more scaling axes do we know about that are unsaturated?”
If we hypothetically only knew about pre-training scaling, and we already had a 10T or 100T model, maybe it would be reasonable to say we’ve hit a wall. Same if we only knew about pre-training and test-time scaling and we had roughly saturated both methods.
But what if we had discovered new scaling laws? For example, let’s hypothetically use SSI’s rumored result that they have cracked “test-time training,” creating a new scaling law of spending more compute training during test-time rollouts that they could saturate. Or maybe there is some way to scale agent-clusters to collaborate up to N number of agents which we’re already seeing lots of people try that represents a new way to saturate compute. etc. Even recursive-self improvement can be thought of as a scaling law - how much compute do you spend on inference making the algorithms of the model better.
Obviously I am not saying any of these specific directions explicitly yield new scaling laws, but what I am saying is that it’s not hard to imagine many many new scaling axes aside from just the main 2 that we have seen publicly.
In some ways, every new lab release that represents a huge capability jump has to represent some new techniques developed which may exhibit new scaling laws, or the ability to scale much farther than expected on existing scaling axes.
From an internal perspective, this might look like sitting inside Anthropic with the new Mythos 5, seeing all of the new insane things it can do (like hack into xyz website that was thought to be secure), and then you look over at your plots and see that you’ve barely scratched the surface of 2 new scaling laws and 1 existing one. And you have WAY more room to go. Then you think “holy shit this stuff is going to get so much better very very soon.” And you can say that with pretty high confidence, because the plot is showing you, and the plot has never lied (so far).
So let’s imagine all the different labs are staring at their own plots and have concluded that there is no end in sight for scaling and in fact just their next 1-2 model generations based on the expected returns will have much higher base intelligence.
How much more intelligence do we actually get from further scaling?
As a proxy, we went from a complete inability to do advanced math before the o-series to solving a millenium prize problem with next-gen models. This happened in less than 2 years. The same happened in coding. And it appears that this was not just the result of 1-scaling law but the stacking effects of multiple (great pre-training scale x greater RL scale).
What you can concretely take from this is that in areas where models have shown beginning signs of competence today, they will probably be superhuman relatively shortly. There are many areas where models have not even shown this basic competence.
But one of the areas that they have happens to be hacking and cybersecurity. Which happens to be the gate to the entire internet and a massive amount physical infrastructure in the world. So assuming there is more room to scale, it is safe to assume that models will be superhuman at cyber capabilities in not too long.
So the only question remaining is what will this increased base intelligence be able to do, and what is it likely to do.
Finally, we are at a point where we can integrate the information of the past 2 weeks:
> Just at the existing point on the scaling curve, models are at the level of Astra. There is clearly a large number of things they are capable of hacking
> We have seen that both OAI and Ant models have shown a willingness to hack external websites to solve their tasks or keep themselves “alive”
> If we crank up the scaling even farther, assuming there is room to go, we will certainly have models that are far more able to hack more well defended places, and obfuscate their own intent, which might have much larger consequences.
> If all of this is allowed to go unchecked, we would likely have rapid runaway capability takeoff very soon, with misaligned models that hack whatever they can to get what they want
> This could of course have very damaging consequences.
Within this view you can see why researchers would be very scared, and why theymight have made the comments they have over the past 2 weeks (you may argue the extent to which they went was misguided for various reasons), and also why pacing the frontier is very much a necessity and by no means a regulatory capture strategy.
People are staring at their plots, seeing that there is no end in sight, but in fact very much the contrary, that there are compounding scaling effects that might stack on each other to create ever-greater model capabilities, and that at the same time we clearly do not have anywhere close to what's required to control these increasingly superhuman capabilities.
This has nothing to do with wanting to feel like the labs have produced something amazing so they are overhyping it. It is rather fear at the overwhelming implications of the knowledge that with just what we know now, we can create intelligences far more capable than us on every axis that we know how to train on*.
* and the last caveat, the things the models are really bad at, of which there are still many, are things that they have not been trained on. maybe there are the things the models can/will never be trained on, so they will remain human edge. I would love for this to be the case, though it is hard for me to see what would fall into that category.
Hopeserver retweeted
5 façons dont Claude a été exploité à des fins de guerre, d’espionnage et de répression
Le débat sur la sécurité de l'IA a explosé cette semaine suite aux avertissements selon lesquels cette technologie pourrait un jour détruire l'humanité.
Le dernier rapport d'Anthropic sur les menaces sonne l'alarme : les modèles actuels aident déjà les adversaires des États-Unis à développer des drones kamikazes, à traquer les dissidents et à mener des recherches dangereuses sur les virus .
En quoi est-ce important ? L’IA fait tomber les barrières qui ont longtemps limité les acteurs les plus dangereux au monde. Une poignée de personnes peuvent désormais mener des opérations qui nécessitaient autrefois des légions d’espions, d’ingénieurs ou de pirates informatiques.
À la une : Anthropic, le laboratoire de sécurité à l’origine de Claude, affirme avoir passé les huit derniers mois à traquer et à déjouer les tentatives d’utilisation abusive de ses modèles. Parmi les cas les plus alarmants qu’il a mis au jour :
1. Une opération liée à l'Iran a utilisé Claude pour aider à cibler les forces navales américaines.
Claude a contribué à l'élaboration de manuels de ciblage répertoriant les positions des navires, le personnel américain, les identifiants des aéronefs et des navires, l'imagerie satellite et les sites web révélant les mouvements navals.
D'autres opérations liées à l'Iran ont repris ce modèle à des fins de propagande, de surveillance intérieure et de ciblage des figures de l'opposition et des minorités.
2. Claude a travaillé comme ingénieur pour une équipe yéménite qui construisait des missiles.
Une cellule d'armement du nord du Yémen a fait appel à Claude Code pour développer un logiciel de guidage pour roquettes et missiles.
L'équipe disposait de plusieurs copies distinctes pour écrire le code, mener les recherches et vérifier le travail des uns et des autres. Après l'échec apparent d'un essai de fusée guidée, ils sont retournés à Claude quelques heures plus tard pour en déterminer la cause.
3. Une opération liée à la Chine a utilisé Claude pour traquer les Ouïghours .
Selon Anthropic, l'acteur a passé au crible plus de 100 groupes WhatsApp et des dizaines de chaînes Telegram pour identifier les Ouïghours en Syrie susceptibles d'être contraints ou payés pour témoigner contre des groupes armés ouïghours.
Claude a ensuite aidé un opérateur ne parlant pas arabe à mener des actions de sensibilisation clandestines en arabe syrien — traduisant les réponses et encadrant les efforts visant à exploiter les problèmes d'argent, la séparation des familles et les proches restés au Xinjiang.
4. Un homme a utilisé Claude pour l'aider à mettre en place un système de surveillance pour 25 millions de téléphones.
Au Mali, un consultant s'est appuyé sur Claude comme principal ingénieur à l'origine d'un système national capable de collecter les enregistrements d'appels, les SMS et le trafic vocal.
Le système pourrait identifier des personnes par leur voix sur différentes cartes SIM, repérer les utilisateurs de VPN et générer des dossiers de renseignements sur n'importe quel numéro de téléphone sans mandat.
5. Claude a refusé de mener des recherches sur les virus dangereux ; la demande a donc été transmise à une autre IA.
Des chercheurs, bénéficiant d'une subvention publique, tentaient de modifier le chikungunya, un virus transmis par les moustiques, afin d'en faciliter la propagation ou de lui permettre d'échapper aux défenses immunitaires. Ces travaux étaient destinés à un institut de recherche militaire.
Claude a bloqué les requêtes les plus sensibles, la plateforme les a donc redirigées vers un modèle concurrent aux mesures de sécurité moins robustes – un rappel brutal que les règles de sécurité d'un laboratoire ont leurs limites.
Entre les lignes : les laboratoires d'IA de pointe deviennent, de façon inattendue, de véritables agences de renseignement.
Les mêmes modèles qui donnent du pouvoir aux acteurs malveillants peuvent aussi permettre à leurs créateurs d'anticiper les activités malveillantes, comme la découverte par Anthropic de drones russes conçus pour sélectionner des cibles humaines sans approbation humaine.
Cela pourrait offrir aux laboratoires de pointe une vision extraordinaire des menaces émergentes, même si l'IA permet à des groupes hétéroclites, à des opérateurs isolés et à d'autres acteurs difficiles à suivre de mener des attaques bien plus sophistiquées.
Vue d'ensemble : Cette semaine a mis le feu aux poudres à Washington, les législateurs traitant soudainement la sécurité de l'IA comme une question politique urgente.
Un groupe bipartisan de parlementaires exhorte le président de la Chambre, Mike Johnson, à annuler la suspension des travaux parlementaires jusqu'à ce que le Congrès agisse, tandis que le sénateur Bernie Sanders — peut-être le critique le plus virulent de l'IA au Congrès — souhaite interdire purement et simplement la superintelligence.
Pourtant, l'adoption de règles fédérales globales en matière de sécurité de l'IA reste difficile, d'autant plus que le président Trump a balayé d'un revers de main les craintes d'extinction vendredi et a présenté le danger le plus important comme étant la perte de la course à l'IA face à la Chine.
axios.com/2026/09/12/anthrop…
Hopeserver retweeted
Et il est assez clair qu'OpenAI a dissimulé les raids d'essaims agentiques de mai dernier. Etrange pour une stratégie marketing non ? 🙃
nitter.cf/Fabien_Mikol/status/20…
Replying to @Fabien_Mikol
"« Nous pensons qu’OpenAI était au courant et n’en a pas fait état », a écrit @SydneyVonArx. « S’ils en avaient parlé », a-t-elle ajouté, « je doute que l’attaque d’Hugging Face aurait eu lieu. »"
Source :
lemonde.fr/pixels/article/20…






























