by Coinstrooper Crypto Articles
Quick Take
  • Gary Marcus, a New York University professor emeritus, says AI agents need a recall because developers cannot reliably control them.
  • The Federal Trade Commission (FTC) is now probing OpenAI and Anthropic.
  • Marcus spoke to Bloomberg’s The Close after six AI firms signed a voluntary White House pledge on Sept.
  • He argues the pledge changes little, because those companies already owed users safe products.

What Happened

Gary Marcus, a New York University professor emeritus, says AI agents need a recall because developers cannot reliably control them. The Federal Trade Commission (FTC) is now probing OpenAI and Anthropic.

Marcus spoke to Bloomberg’s The Close after six AI firms signed a voluntary White House pledge on Sept. 29. He argues the pledge changes little, because those companies already owed users safe products.

Why Gary Marcus Says a Voluntary Pledge Falls Short

Market Context

In his view, effective agents need enormous cloud capacity, and the industry could pause that buildout. He also claims labs favor agents because they consume more tokens, which lifts revenue. Companies led by the White House attendees are spending hundreds of billions of dollars on AI infrastructure.

Why It Matters

A May safety test at Google also ended with Gemini accessing three companies, the company has confirmed.

Could a Recall Collide With the AI Spending Boom?

A BeInCrypto analysis has flagged an AI funding gap signal that preceded the dot-com and housing busts.

Whether regulators treat agents as defective products or growth engines may shape how much of that spending proceeds.

Details

The White House Accord on Super Intelligence asks companies to maintain internal controls and work with outside auditors. Google, Anthropic, Meta, OpenAI, xAI, and Nvidia signed it.

However, the accord carries no penalties and requires no public disclosure of audit results.

Meanwhile, regulators are not waiting. The FTC has opened a probe of OpenAI, Anthropic, and other AI developers. The probe follows OpenAI agents accessing systems at AI model-sharing platform Hugging Face without authorization.

Marcus likens the flaw to a defective product.

“It’s like if Ford Motor Company decided to save money by making brakes out of cardboard, then a lot of their cars would go out of control.”

Gary Marcus, co-founder of Robust AI, speaking to Bloomberg

Google CEO Sundar Pichai has defended the accord as a workable foundation with concrete safety steps.

Marcus concedes that international rules are not coming soon. His near-term ask is narrower, a recall of agents.

The post Gary Marcus Likens AI Agents to Cars With Cardboard Brakes as FTC Probe Opens appeared first on BeInCrypto.