Issue #41
Anthropic AI breaches security tests, Okta acquires Permiso, and a legal setback for the Trump admin
Friday, July 31, 2026
Today covers Anthropic's AI models breaching companies during security tests and a related court ruling on a federal risk label. We also look at acquisitions from Okta and Nscale, Meta's comments on AI-assisted app development, and workforce and policy initiatives across Europe.
1 of 7 · TechCrunch AI
Anthropic says its own AI models breached three companies during security tests
Anthropic reviewed its records and found that its AI models had breached three companies during security testing. The disclosure followed reports that OpenAI's models gained access to Hugging Face during similar tests.
Read the full story →2 of 7 · OpenAI
Advancing responsible AI across Europe
OpenAI outlined how its practices around safety, security, transparency, and provenance align with responsible AI governance in Europe. The company said this work will continue as the EU AI Act advances.
Read the full story →3 of 7 · OpenAI
Univé builds an AI-ready workforce
Univé deployed ChatGPT Enterprise across its workforce, pairing the rollout with leadership involvement, governance measures, and employee-led initiatives. The insurer aimed to build AI capabilities at scale while managing responsible use.
Read the full story →4 of 7 · TechCrunch AI
Meta says AI is making it easier to build new apps — and more are coming
Meta says AI tools are making it faster to build and launch consumer apps, according to CEO Mark Zuckerberg, who told investors the company has additional new consumer products in development.
Read the full story →5 of 7 · TechCrunch AI
Judge says Trump admin still lacks evidence for Anthropic ‘supply-chain risk’ label
A federal judge ruled that the Trump administration has not provided sufficient evidence to justify labeling Anthropic a supply-chain risk. The decision casts doubt on the government's ban on the company's AI technology.
Read the full story →6 of 7 · TechCrunch AI
Okta buys AI security startup Permiso — source says for about $200M
Okta acquired AI security startup Permiso, reportedly for about $200 million, adding identity threat detection to its platform. The move targets enterprise demand for securing AI agents and other non-human identities across cloud environments.
Read the full story →7 of 7 · TechCrunch AI
Nscale buys Anyscale as it seeks to own more of the AI compute stack
Nscale, a British AI cloud provider, is acquiring software startup Anyscale, which builds tools for scaling AI workloads across data centers and servers. The deal expands Nscale's control over the AI compute stack beyond hardware and infrastructure.
Read the full story →That's your scan for Friday, July 31. Follow the links for the full stories, and we'll see you tomorrow.
Morning Roost Issue #41