June 10, 2026

BT Taps Anthropic’s Powerful Project Glasswing to Harden UK Network Security and Secure Customer Engagement

BT has confirmed it is joining Anthropic’s Project Glasswing, becoming the first UK enterprise to publicly declare membership of the initiative. The announcement came at the UK Government’s AI Adoption Summit, where BT Chief Executive Allison Kirkby committed the company to deploying Anthropic’s Claude Mythos Preview, an unreleased frontier model, across its cybersecurity defence operations.

Project Glasswing launched in April 2026 with a straightforward premise. Anthropic had built a model capable of finding and exploiting software vulnerabilities at a level surpassing most human security researchers. Anthropic stressed that the responsible move was to put it to work on defence before others used equivalent tools for attack. The model remains unavailable to the general public for precisely that reason. Claude Mythos Preview has already identified thousands of high-severity vulnerabilities across all major operating systems and web browsers.

Kirkby commented:

“AI only works at scale when it is underpinned by future-ready networks that are secure, resilient, safe.”

The initial cohort of Glasswing partners read like a who’s-who of global tech infrastructure. Amazon Web Services, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, Microsoft, and NVIDIA were among the first organisations given access, alongside critical software maintainers. BT is the first UK-headquartered company to confirm it is on the list.

On 2 June, Anthropic expanded the programme to approximately 150 new organisations across more than 15 countries. Participants span telecoms, energy, healthcare, and government. Partners have collectively identified more than 10,000 high- or critical-severity vulnerabilities across the programme.

BT Business CEO Jon James was direct about the company’s motivation:

“By joining Project Glasswing, BT will strengthen its own cyber security capability to protect our networks, our customers and the wider UK.”

Why BT’s Deal With Anthropic and Project Glasswing Points to a Market Moving Quickly

The timing of BT’s announcement reflects the changing winds in how tech vendors are approaching cybersecurity. AI-assisted attacks, which are faster to deploy, harder to detect, and increasingly automated, have pushed traditional signature-based defences to their limits. According to Bridewell’s Cyber Security in CNI Report 2026, half of critical national infrastructure organisations had already experienced IT outages or operational disruption from cyberattacks. Nearly a third report financial losses.

Anthropic’s read on the market is that this window is critical. The company expects many other AI providers to have developed models as capable as Mythos Preview within six to twelve months. Anthropic suggests that some of these may be released without the safeguards built into Project Glasswing. OpenAI has already launched GPT-5.5-Cyber, its own cybersecurity-focused model. This is currently being tested with a group of partners. The race to establish responsible norms before capable models proliferate is, at least in part, what Glasswing is about.

What Does the BT Deal Mean for CX, IT and Cybersecurity Leaders?

Organisations running customer-facing operations on BT infrastructure, such as contact centres, cloud communications platforms, and CRM integrations, face several implications. Vulnerabilities in the underlying network can now be identified and patched before they are exploited, rather than after. That matters because a successful attack on shared telecoms infrastructure does not stay inside an IT department. It disrupts agent availability, degrades platform performance, and surfaces as a customer experience problem.

Alongside Mythos Preview access, Anthropic is providing partners with an automated scanning and reporting framework and a threat-modelling tool for prioritising attack targets. This gives BT’s security teams a more systematic approach to vulnerability management than reactive monitoring alone can provide. Separately, the Government confirmed BT will share data on its workplace AI usage to inform broader enterprise rollouts. This suggests that the Anthropic relationship has a workforce dimension too, not just a defensive one. CX leaders and enterprise buyers evaluating managed connectivity or infrastructure services would do well to ask suppliers directly how AI-led vulnerability management features in their security operations.

Customer Experience Digital Experience