Report from the Vatican - Marcus on AI - Substack
In my view, Pope Leo XIV is by far the clearest thinker on AI among prominent world leaders. ... Such an approach risks obscuring human responsibility ...
Marcus on AI - Substack 36
Covers: AI Safety & Existential RiskChatbots & LLMsHallucinations & FabricationRegulation & PolicyAI Agents & AutomationGovernment & Public SectorBias & FairnessMilitary & DefenseChina & GeopoliticsCybersecurity
Will Generative AI do more harm than good? - by Gary Marcus
We have nonconsensual deep fake porn, a rise in phishing scams, wholesale disinformation, cognitive surrender, threats to democracy from the subtle ...
Chatbots & LLMsDeepfakes & Synthetic MediaMisinformationCybersecurity
We must recall open-ended AI agents with internet access from the market, now
I think OpenAI and its peers are now producing a technology that is more capable and poses more risk than what was being made even six months ago. I'm ...
Coming soon: New York City's hearing on AI risks - Marcus on AI
My name is Gary Marcus. I am a scientist, author, and NYU Professor Emeritus who cofounded an AI startup that was acquired by Uber. My Substack, ...
BREAKING: Two rays of hope - by Gary Marcus
As I have argued here multiple times, the current lack of transparency in U.S. AI ... risks — yet the government apparently waved the model through.
No, Anderson Cooper, AI is not going to kill all humans by 2030 - Marcus on AI
AI is already causing harms, there are are many real risks , but constant focus on absurd end-of-the-world fantasies has made the situation worse.
The Case for Boycotting Generative AI - by Gary Marcus
They have no serious plan for mitigating the risks of generative AI . ... But the risk of catastrophe is real, e.g., from AI -generated pathogens ...
Red Alert: OpenAI is poised to cross an AI safety redline. - Marcus on AI - Substack
performance gain feels like a dangerous game. Earlier tonight Steven Adler, of Guidelight. ai and one of the many researchers to have departed from ...
The new Sanders-Casar Ban Artificial Superintelligence Act – and why I oppose it
Furthermore, it focuses too much on hypothetical future risks (such as AI that “plan[s] ... to disempower humanity) to the exclusion of current risks , ...
Dwarkesh Patels's wildly popular but dangerously misleading account of the OpenAI ...
They are extremely biased towards the reality they want rather than the reality we live in. ... Just as boys must be boys, AI must be AI . We will even ...
5 lessons from the OpenAI / Hugging Face incident - Marcus on AI - Substack
... threats posed by AI . On Wednesday, METR released a (partly) ... AI that is vulnerable to these risks in the first place. Society ought ...
Excellent new Bill Gates essay on the urgency of having a coherent AI plan
The risks that AI creates, around bioterrorism, deepfakes, disinformation, cyberattacks, and so on, empowering criminals such that “Even criminals ...
Regulation & PolicyDeepfakes & Synthetic MediaMisinformationCybersecurity
An open letter to David Sacks - Marcus on AI - Substack
China is closing the gap, but it's not because we have too much regulation on AI .
China catches up - Marcus on AI - Substack
“America's preoccupation with “winning” the AI race with China could well lead to unprecedented catastrophes, even catastrophes on a global scale. Not ...
The White House's shambolic AI policy - by Gary Marcus
For now, and until the US government takes a more comprehensive approach to addressing AI risks , states will have little choice but to file ...
You can't get more 2026 than that - Marcus on AI
Amazing: KPMG wrote a report describing the successful use of AI by businesses. But the case studies turned out to be AI hallucinations .
Breaking: Google liable for hallucinations - Marcus on AI - Substack
Why do AI companies get a pass on being responsible for the dangerous malfunctioning of their products when operated perfectly normally? I have a ...
Breaking: Google liable for hallucinations - Marcus on AI - Substack
Sorry to bother your mailbox again but this legal decision is potentially huge, especially if it spreads and other countries make similar ...
The revenge of Claude Mythos - Marcus on AI
and how Tom Friedman freaked out about it at the New York Times? And how Anthopic said Mythos was too dangerous to release, and held hushed meetings ...
The illusion of Generative AI , the insanity of massive bets on hyperscaling, and the case for ...
Excellent brief but focused interview on the risks of massive bets on hyperscaling, with Zachary Karabell; keynote earlier this week at Web Summit:.
Breaking: Autonomous Agents are a Shitshow - by Gary Marcus
... AI Labs, examining 847 autonomous agent deployments ... 91% were vulnerable to subtle but dangerous tool-chaining attacks (see diagram below),.
Dario Amodei, hype, AI safety, and the explosion of vibe-coded AI disasters - Marcus on AI
Dario Amodei, hype, AI safety, and the explosion of vibe-coded AI disasters ... classic attribution bias ... when the AI succeeds, it's a “miracle of ...
What should we take from Anthropic's (possibly) terrifying new report on Mythos? - Marcus on AI
A system doesn't have to be AGI to carry risks . The strongest lesson here is about policy. Anthropic showed some admirable restraint in not ...
Is the US military actually afraid of Claude? A new theory of why Anthropic was labeled a ...
Marcus on AI . Subscribe Sign in. Is the US ... And there is no argument Anthropic's system of guardrails is any more dangerous than anyone else's.
Breaking: “sycophantic AI distorts belief, manufacturing certainty where there should be doubt”
And as the article says, such AI can “facilitate delusion-like epistemic states, producing belief markedly divergent from reality.” The paper, which ...
Code Red for Humanity? - Marcus on AI
We are on a collision course with catastrophe. Paraphrasing a button that I used to wear as a teenager, one hallucination could ruin your whole planet ...
Rumors of AGI's arrival have been greatly exaggerated - Marcus on AI
AI Risk Management Framework. https://www.nist.gov/itl/ ai - risk -management-framework. Nudo, J., Pandolfo, M. E., Loru, E., Samory, M., Cinelli, M ...
Sir Demis Hassabis becomes the latest to say that ChatGPT is a dead-end and that we must ...
... AI and world (aka cognitive) models. Generative AI's crippling and widespread failure to induce robust models of the world · Gary Marcus. ·. June 28 ...
The (possibly) coming AI backlash and information warfare - Marcus on AI - Substack
So will his utter failure to meaningfully regulate AI , against the will of voters and political leaders both left and right. … Generative AI , once ...
Three years on, ChatGPT still isn't what it was cracked up to be – and it probably never will be
AI with the breadth and resourcesfulness of human intelligence) is not ... hallucinate ; they still can't match domain specific models, the ...