Welcome to the latest edition of the AI Update, where we continue to track the major developments in artificial intelligence and their implications for identity security and cybersecurity more broadly.
The White House has asked OpenAI to stagger the release of its next AI model, and will approve access to it on a customer-by-customer basis before a wider rollout, according to new reports. It’s the latest sign that we really are heading toward some kind of nationalization of frontier AI labs, after the White House imposed export controls on OpenAI rival Anthropic’s big new model a couple of weeks ago. These moves are ostensibly about AI safety, with the new models reportedly being very good at cybersecurity, both on defence and offence. In a memo to OpenAI staff, CEO Sam Altman said, “We’ve made clear to the U.S. government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases.”
Anthropic has formally accused Alibaba and affiliates of its Qwen AI lab of a large-scale “distillation attack” on its Claude models. This is effectively an accusation of IP theft: distillation is the process of running prompts on a model at scale and trying to reverse-engineer it based on its responses. Anthropic laid out the accusation in a letter to U.S. Senators Tim Scott and Elizabeth Warren ahead of a Senate Banking Committee hearing, asserting that Alibaba sent roughly 28.8 million exchanges with Claude through nearly 25,000 fraudulent accounts between April 22 and June 5, 2026.
The Five Eyes intelligence network has issued a warning that AI-fuelled cyberattacks will exceed expectations – and many defensive capabilities – on a timeline of “months, not years.” Its joint statement urged business and government leaders to prioritize cyber defences immediately, and was signed by senior cyber officials from the Australian Cyber Security Centre, Canadian Centre for Cyber Security, New Zealand’s National Cyber Security Centre, the U.K.’s National Cyber Security Centre, the U.S. National Security Agency, and CISA.
Hyundai Motor workers in South Korea have voted to strike, in part over concerns about being replaced by robots. Talks with management have mainly revolved around wages, but the union also wants more of a say in how robots will be deployed after Hyundai announced plans to deploy Boston Dynamics’ Atlas humanoid robots at its Georgia plant starting in 2028. “Remember that without labour–management agreement, not a single robot using new technology will be allowed to enter the workplace,” the union said.
Santa Clara-based Cequence Security has launched a biometric CAPTCHA system for the AI agent era. Its Intent Graph and Biometric Check system is meant to distinguish human users from bots, replacing CAPTCHAs, SMS codes, and the like with hardware-bound cryptographic attestation through a device’s Secure Enclave. If the system detects a potential agentic bot, it prompts biometric authentication via Touch ID, Face ID, or Windows Hello.
The Linux Foundation has announced the Agent Name Service, or ANS, a planned open standard that would essentially let people use the Domain Name System to register AI agents in a similar way to how websites are registered via DNS. The system would let users verify who a given agent represents, its permissions, and its operational history. ANS will anchor agent identity records directly to DNS. “By building on proven internet foundations, ANS creates a path for agents to be identified and discovered across the open web, helping ensure the next era of innovation remains as open and interoperable as the last,” said Jared Sine, the Chief Strategy and Legal Officer of domain registration service GoDaddy.
The chatbot’s take: We got a little creative this week.
–
By Alex Perala
Follow Us