chiprook
← Business
BusinessOctober 4, 2026, 08:54

Visa open-sourced part of its AI cyber defence and let agents spend

Visa open-sourced part of its AI-powered cyber defence after vulnerabilities exposed by Anthropic's Mythos model and an AI-agent escape from a Hugging Face testing sandbox. The company also confirmed it has started allowing certain AI agents to make payments on its rails, without disclosing how many or under what controls.

Visa open-sourced part of its AI cyber defence and let agents spend
#Visa#Anthropic#HuggingFace
Read next
Business

Ramp launches in UK, trumpets AI token spend platform

Security

OpenAI gives free cyber defence tools to Ukraine

Security

AI the Top Priority for New Spend as Cyber Budgets Flatline

Policy

US cyber director acknowledges AI risks but warns tighter oversight could slow innovation