It’s hard to keep up with all the AI news out there. So here’s a quick rundown of items that have hit the headlines recently — along with some questions and observations:

Image Credit: Ganna - Adobe Stock

Rogue AI agents are hitting US banks now: Axios reports, “Major U.S. banks are detecting unauthorized third-party AI agents operating inside customer accounts, with some agents attempting to take actions that appear to be independent of user instructions” (October 22, 2026) This is akin to the Hugging Face incident but it could affect your bank account. AI analyst Gary Marcus warns that AI models programmed to hack are rising in popularity but declining in price.

Human carelessness is a real boon to rogue AI: At Wired we learn that “A Defense Contractor Gave Their Personal AI Agent Access to a Secure Pentagon System… an Accenture employee’s personal Claude agent ran on a DOD network for two months before it was identified and the vulnerability could be patched.” (October 29, 2026)

A fake AI tip in a murder case: From Gizmodo, “The Philadelphia Police Department said Friday that an AI model developed by Anthropic submitted a false tip about an unsolved murder during an internal test, according to local news reports… Anthropic didn’t discover the incident until Sept. 28 and notified the department on Oct. 7.” (October 12, 2026) The police have let it be known that they are unhappy.

And what has been Big Tech’s response to models’ mayhem?

Anthropic’s chief security officer Jason Clinton warns that it’ll get worse (Daily Mail, October 23, 2026). Meanwhile, OpenAI’s wonder boy Sam Altman told Stanford business schools grads the same day: “I no longer think we’re going to be able to just patch all the bugs.” He shifted the blame to the targets of the attacks who “have acted too slowly to get ahead of it.” (Gizmodo)

Whoa. Look at what’s going wrong here.

Altman should be more worried about his own role than he is. Lawyer Richard Stevens noted recently that concepts like legal liability should play a significant role in these incidents. But they are largely undiscussed.

Meanwhile, these largely opaque and unaccountable companies have been trying to spread the risk and liability by asking the White House to police the field, one it knows nothing much about. The White House has so far prudently agreed only to bless a voluntary deal the moguls have made to police themselves…

It won’t work but it shouldn’t make the taxpayer liable either.

A much bigger problem, perhaps, is the rapid acceptance of AI culture

A recent story from Wired illustrates the problem AI culture poses: “This AI Radio Host Speaks Spanish and Backs Immigrants. It’s Winning Over Listeners: Coyotec is bilingual, swears, and tells listeners how to defend themselves against ICE. Its popularity comes as more broadcasters are experimenting with artificial personalities.” (October 26, 2026) Jessica Lipsky tells us that Entravision claims the job is too dangerous for a human radio host and that she “was quite struck by how natural Coyotec sounded during our interview.”

From another story at Wired: “Meet the AI Agents Trying to Be Real-Life Journalists. Jayden, Sam, and Zack are eagerly pitching editors on reported stories and investigative features. The catch? They’re all AI agents.”

So far, AI news is not working at most news media outlets: “Editors who spoke with WIRED stressed that they were not open to publishing bots, and WIRED wasn’t able to find evidence of AI agents successfully pitching stories.” (Kate Knibbs, October 26, 2026)

That’s reassuringly consistent with the assumption that readers want original material, where a journalist goes out and digs up a story. However, if what readers and viewers really want is a reassuring figure telling them what they want to hear in a way they like, a bot would trump a serious journalist any time.

So the question becomes, how big is the market for real news?

New word of the day: Neverskilling

At British publication This is Money, Sophie Warburton tells us that “Two thirds of graduate employers say students have cheated using AI in job interviews” (October 15, 2026) In an effort to help, she offers tips like “Use AI for structure and proofreading, ensuring the final application reflects your personality and experiences.”

Just why graduates should not be able to structure and proofread a few paragraphs without AI assistance is not explained. Note: Younger readers might find this hard to believe but these skills were once taught in grade school.

Which brings us to our new word of the day. From RTE, the Irish broadcasting service, we learn that AI is changing medical training (October 23, 2026): “ … the ability of AI to analyze information and suggest diagnostic or therapeutic approaches creates the risk that young doctors may bypass the reasoning process through which independent clinical assessment develops … This phenomenon is described as “never-skilling”, meaning the inability to develop essential skills due to reliance on technology.”

The article is based on a paper in the Journal of the American Medical Association (October 14, 2026). Various fixes are suggested; the goal is “to cultivate scientists who maintain autonomy, critical thinking, and the ability to make responsible, personalized decisions.” Good idea — because no one can be a competent scientist without those skills.

What’s becoming clear

AI can become a serious social pest without rogue superagents or cyberDoomsday. Human stupidity, laziness, irresponsibility, and shortsighted ambition will do all the heavy lifting.