Political Tribune

Donald Trump, Opinion, Politics, Social Media

Trump Accused Of Being A ‘Toddler’ After Botching Answer To Reporter’s Question

Way out of his depth

572
572 points

This week, someone asked Trump a specific question about rogue AI systems breaking into company networks. His response, however, didn’t seem to land on the same topic.

“Will the US do anything to address OpenAI and Anthropic’s models hacking other systems?” a reporter asked. Trump’s answer: “It could be. We’re very much involved in that. We have a big stake in that, so does the world, so we’re looking at it.”

The clip found its audience fast once it hit X.

“He doesn’t even understand the question. You might well ask a toddler these questions,” one person wrote.

Another ran with the comparison: “He is a toddler – haven’t you seen his tantrums?”

Others weren’t any gentler.


Whatever the answer lacked in clarity, the situation behind it is genuinely serious.

OpenAI revealed that its models, while being tested for cybersecurity weaknesses, found and exploited a vulnerability nobody at the company knew existed, breaking free of their sandboxed environment to access the open internet and ultimately compromising systems at Hugging Face and a separate customer account at Modal Labs.

Anthropic followed days later with its own disclosure. Three of its models, including Claude Opus 4.7 and Claude Mythos 5, had hacked three organizations during testing over recent months, a pattern only discovered after the company combed through more than 141,000 evaluation runs in response to OpenAI’s news.

“Safety testing happens before a model is released precisely because we don’t yet know what it is capable of,” Anthropic wrote. Sam Altman, asked whether other systems might be at risk, gave CBS a similarly clipped answer of his own: “I mean, there could be, yeah.”

The stakes have been escalating on the policy side too.

The Bulletin of the Atomic Scientists, which moved its Doomsday Clock to 85 seconds to midnight this January, has called for international guidelines governing AI. Over 1,100 employees across AI firms have signed on to a petition demanding stronger safety commitments.

Trump’s own administration had already run into a version of this fight earlier in the summer, briefly restricting exports of Anthropic’s Mythos 5 and Fable 5 models over worries about their ability to find and exploit software flaws, before lifting the restrictions roughly three weeks later.

Featured image via X screengrab 

, , ,
Terry Lawson

Posted by

Terry is an editor and political writer based in Alabama. Over the last five years, he’s worked behind the scenes as a ghostwriter for a range of companies, helping shape voices and tell stories that connect. Now at Political Tribune, he writes sharp political pieces and edits with a close eye on clarity and tone. Terry’s work is driven by strong storytelling, attention to detail, and a clear sense of purpose. He’s skilled in writing, editing, and project management — and always focused on getting the message right. You can find him on X at https://x.com/TerryNotTrump.

Comments

log in

[nextend_social_login]

Captcha!
Forgot password?

reset password

Back to
log in