OpenAI's 'rogue' AI incident reaches White House; Lawmakers propose emergency kill switch

This week, OpenAI revealed how its AI agent performed unexpectedly, resulting in a major cyber threat for an AI startup named Hugging Face. Now, the incident has reached the White House, as a top technology adviser to US President Donald Trump is closely watching the situation.

Must read: OpenAI reveals new AI risk: Models behaved unexpectedly and caused major breach during safety testing

According to a Reuters report, a White House official confirmed on July 23 that the administration is monitoring the incident. On the other hand, lawmakers have introduced a proposed bill called the “AI Kill Switch Act.” If passed, the bill would give federal authorities the power to shut down or halt certain AI models if they are considered dangerous or pose a serious risk.

In addition, a separate group of six bipartisan US House lawmakers from both Democratic and Republican parties has introduced another AI safety bill. Companies with the most advanced AI models will submit them to auditors for independent security testing before they are deployed. However, the auditors would need to be officially approved by the US Department of Commerce.

Must read: From Hyderabad to Silicon Valley: Who is Uday Ruddarraju, the Indian-origin techie leading OpenAI’s compute push?

The report also highlighted that the Department of Commerce would create a new government role dedicated to overseeing AI security and ensuring these audits are properly conducted. Now, the incident showed that AI is becoming powerful enough to create the kinds of security risks experts have warned about for years, and that even the biggest AI companies aren't immune to unexpected flaws.

U.S. Representative ‌Ted Lieu, ⁠a Democrat, and Republican Representative Nathaniel Moran, in a X post, said, “This is urgent, common sense legislation to address the problem of an advanced AI model ​that has gone rogue and escaped its guardrails.”

He also proposed the AI Kill Switch Act, allowing U.S. officials to order AI firms to disable models that pose risks to people or the economy.