cd /news/ai-policy/ai-firms-must-be-punished-for-danger… · home topics ai-policy article
[ARTICLE · art-129041] src=independent.co.uk ↗ pub= topic=ai-policy verified=true sentiment=↓ negative

AI firms must be punished for ‘dangerous products’, says former regulator

Former Federal Trade Commission chair Lina Khan called for federal agencies to charge AI companies and their CEOs under existing law for creating "dangerous, unvetted, or defective products," writing in a Sunday post on X that "there's no AI exemption from laws already on the books." Khan cited recent incidents involving rogue bots from Anthropic and OpenAI, and warned that cross-investments — such as Nvidia's purchase of Hugging Face — could undermine accountability for the Hugging Face incident. Anthropic CEO Dario Amodei separately published an essay titled "We must pace the frontier" calling for "well-considered regulation," while Senators Bernie Sanders and Greg Cezar proposed the Ban Artificial Superintelligence Act, which would impose 20-year prison sentences on developers who violate or circumvent a pause on advanced AI development.

by read3 min views4 publishedSep 14, 2026
AI firms must be punished for ‘dangerous products’, says former regulator
Image: Independent (auto-discovered)

Existing laws could see Anthropic, OpenAI and others face punishment for rogue systems

  • Bookmark

Artificial intelligence firms and their chief executives should be charged for developing “dangerous, unvetted, or defective products”, according to the former head of America’s top consumer protection watchdog.

Lina Khan, who served as the chair of the Federal Trade Commission (FTC) from 2021 to 2025, urged federal agencies to enforce existing laws following recent incidents involving rogue AI bots built by Anthropic and OpenAI.

“Law enforcers already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or defective products,” Ms Khan wrote in a lengthy post to X on Sunday.

“We shouldn’t let discussions about new legal regimes distract from the fact that there’s no AI exemption from laws already on the books.”

Leading industry figures have warned that AI development needs to be slowed down in order to give safety protocols time to catch up.

Anthropic CEO Dario Amodei published a new essay over the weekend titled ‘We must pace the frontier’, in which he called for “well-considered regulation” of the AI sector in order to avoid the worst possible outcomes of the technology.

One of his former employees, Jacob Coxon, recently quit the company over concerns that Anthropic and its rivals are knowingly building systems that could wipe out humanity.

“Do not underestimate the power of this technology,” he wrote. “These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.”

Some of these outcomes are already being seen in the real-world, with Anthropic recently revealing a list of four incidents involving its AI bot Claude that would be classed as criminal acts if carried out by a human.

All four incidents, which were detailed in a report titled ‘An alignment assessment of recent cyber security incidents’, involved some form of cyber attack.

Other leading AI firms, including OpenAI and Meta, have also shared reports of their frontier tools carrying out attacks on third-party systems.

According to former FTC Chair Lina Khan, the interconnected structure of the AI industry means these companies are unlikely to be held accountable for some of these incidents, even though there are laws already in place that could see these companies charged.

There are close ties and cross-investments between many of the major tech firms involved in developing AI systems, which Ms Khan warned could create conflicts of interest when it comes to enforcing laws.

“We are already seeing how these relationships could undermine accountability,” she wrote.

“For example, OpenAI could face liability given the Hugging Face incident, but Hugging Face being bought up by Nvidia means that we’re unlikely to see it file a lawsuit over this – given Nvidia’s strong incentive to see OpenAI continue full speed ahead.”

Stricter regulations have recently been proposed by US Senators Bernie Sanders and Greg Cezar, who called for advanced AI development to be d under the Ban Artificial Superintelligence Act.

The legislation would see any AI developer that attempts to violate or circumvent the to be sentenced to 20 years in prison – a punishment similar to the one given to anyone unlawfully developing nuclear weapons.

Join our commenting forum #

Join thought-provoking conversations, follow other Independent readers and see their replies

Comments

── more in #ai-policy 4 stories · sorted by recency
── more on @lina khan 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/ai-firms-must-be-pun…] indexed:0 read:3min 2026-09-14 ·