All insights

Weekly16 October 20261 min read

This week in AI: AI on the defenders' side

Anthropic launched a cyber mission and widened its verification programme, and Google is testing a new model for cyber defence. For most organisations, the basics still matter most.

What happened

On 6 October, Anthropic expanded its Cyber Verification Program, and on 8 October it introduced the Anthropic Cyber Mission, focused on helping defenders. Google said Gemini 4 Carbon, a model for long-running engineering and cyber defence work, is in pre-release testing with trusted cyber defenders. Google also added an opt-in Android feature that uses AI to spot conversational scams in messages.

Our take

AI is becoming good at finding weaknesses in software. That cuts both ways, so it is encouraging to see labs putting effort into giving defenders an advantage.

For councils, housing associations and charities, though, the most effective protections are still the basics. Keep systems patched. Turn on multi-factor authentication everywhere. Limit who has administrator access. Keep tested backups. Train staff to spot phishing, which AI is making more convincing. Cyber Essentials covers much of this, and we are working towards it ourselves.

AI can help with the basics too: summarising security alerts, spotting unusual activity and checking configurations. As with any AI tool, it works best when a person reviews what it finds and decides what to do.

Scam protection for the public is a welcome development. Many of the people public services care for most are the ones scammers target.

Keep reading

Related insights