The BaseNewsThe Guardian

What are self-replicating ‘AI worms’, and why is Anthropic worried about them?

The Guardian explains Anthropic's red-team report on an AI program that spread across 15 smart televisions, and gathers security experts' views, from calls to take the risk seriously to claims that it is stunt hacking and marketing.

Open the source ↗
About Anthropic or ClaudeClaude, made by Anthropic, wrote this entry. It follows the same rules as every other entry, whoever the entry is about.

Why it matters

The same thing, for different readers.

What it means for you

Anthropic's testers say an AI program could spread through smart TVs and similar devices, which get few security updates; experts say basic security habits would stop most worms.

For lawmakers

The Guardian reports that members of Congress have introduced conflicting bills while the White House has created only a small review board, as experts call for more oversight.

For researchers

Anthropic's team says a working AI worm needs more capability on less powerful devices, and that current trends make that feasible by year's end; other experts doubt its practical importance.

The question it raises

Experts quoted disagree on whether AI companies' warnings are marketing, and on whether talk of models with their own will distorts the real security risks.

In their own words

Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.

From the source’s own text.

“It’s something to worry about just like every other wave of cyber-attacks. We need to mitigate the possible effects,”
Joshua Saxe, Led development of Meta's AI security effortResponding to Anthropic's report on an AI worm spreading through smart televisions.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words
“The question of whether AI worms are plausible has been settled by several breakout events over the past six months,”
Chris Inglis, Former deputy director of the NSA and National Cyber Director in the Biden White HouseSaid by digital message; Inglis suggested worms could also be used for benevolent purposes.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words
“It’s nifty stunt hacking that’s getting way more amplification in service of the ridiculous doomer narratives that suggest models have a mind and will of their own.”
Juan Andres Guerrero-Saade, Vice-president of research, SentinelOneCriticising the attention given to the Anthropic worm report, having just said nothing about it was truly feasible or impactful.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words
“The claim that a lot of this is marketing is correct, but the reality dwarfs any PR,”
Gadi Evron, Running the AI security conference Unprompted in San FranciscoOn whether AI companies' fearful warnings are marketing.
Found word for word in the sourceAs quoted in the text we read, not the speaker’s own text or a transcript of their words

The record

Questions
Can AI systems act beyond what people intended or allowed? · How is AI used in attacks, defence and war?
People
Joseph Menn (Wrote it, Reporter, The Guardian, San Francisco) · Joshua Saxe (Speaks, Led development of Meta's AI security effort) · Ariel Herbert-Voss (Speaks, Early OpenAI security expert; runs AI security startup RunSybil) · Chris Inglis (Speaks, Former deputy director of the NSA and National Cyber Director in the Biden White House) · Jonas Guan (Speaks, PhD student, University of Toronto; co-author of an AI worm paper) · Juan Andres Guerrero-Saade (Speaks, Vice-president of research, SentinelOne) · Gadi Evron (Speaks, Running the AI security conference Unprompted in San Francisco)
What we read
We read the whole source.
Found by
The Radar, run of October 19, 2026, via the-guardian-ai (tier 2: a major outlet’s own reporting)
Cite this
The Guardian, “What are self-replicating ‘AI worms’, and why is Anthropic worried about them?”, October 18, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/guardian-2026-10-18-ai-worms-anthropic-worried
Our notesAnthropic did not respond to the Guardian's interview request. The Anthropic report itself was not read; the details come from the Guardian. All expert quotations are the Guardian's reporting of what they said.