The BaseNewsAxiosOctober 21, 2026

AI is getting better at hacking power and water systems, new research shows

According to the description, Axios reports research by the cybersecurity company Frenos in which 10 AI models were tested on planning attacks against industrial systems, and three of the 10 produced attack plans in roughly 90% of attempts. The full article could not be read.

Open the source ↗

Why it matters

The same thing, for different readers.

What it means for you

Axios reports that AI models can now plan attacks on systems such as water treatment, so operators of such services face growing pressure to defend them.

For researchers

The Frenos tests, as summarised by Axios, measure attack planning rather than execution, so the real-world risk is not settled by these figures.

How the finding was reached

In our words, from the source. Weigh the finding by how it was made.

What was done: Frenos researchers tested 10 AI models on planning attacks against industrial systems, such as software that monitors water treatment.
What it found: Per the Axios excerpt, three of 10 models built attack plans in roughly 90% of attempts; three models seen as frontier models succeeded in 17.9% to 31.9% of attempts; the strongest foreign open-weight model succeeded 36.2% of the time.
What it does not show: Not stated in the excerpt; the full article and the Frenos report were not read.
Status: The company’s own report about its own work

The record

Questions: How is AI used in attacks, defence and war? · How fast is AI advancing, and where is it heading?
People: Sam Sabin (Wrote it) · Steve Swick (Mentioned, former top security executive at American Electric Power)
What we read: We have read only the description the publisher gives. No quotations until someone reads the source itself.
Found by: The Radar, run of October 22, 2026, via axios (tier 2: a major outlet’s own reporting)
Cite this: Axios, “AI is getting better at hacking power and water systems, new research shows”, October 21, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/axios-2026-10-21-ai-hacking-power-water

Our notesFetch returned 403 and the Wayback Machine returned 404; no source text was saved, so the file named in textFile does not exist. Summary and numbers come from the triage excerpt, which is cut off. Check the Frenos report for methods and for which models were tested.