Ethical Hacking News
Anthropic's recent expansion of its Project Glasswing program has raised concerns over cybersecurity safeguards in place for its AI-powered platform. The company's decision to exclude certain UK banks from access to Glasswing has sparked debate over who will ultimately control access to these powerful AI systems, and whether the benefits outweigh the risks. Will Anthropic be able to deliver on its promise of making all software more secure through the use of AI-powered tools like Glasswing?
Anthropic has expanded its Project Glasswing program to include 150 new organizations worldwide, sparking concerns about cybersecurity safeguards. The expansion comes as Anthropic's AI model Mythos is unveiled, touted as an expert bug hunter and zero-day specialist. Anthropic excludes certain UK banks from access to Glasswing, raising eyebrows and concerns about "cybersecurity gatekeeping." Experts warn that limiting access to Glasswing could leave banks more exposed to cyber threats and impact vulnerability management capabilities. Anthropic remains committed to its goal of making all software more secure through the use of AI-powered tools like Glasswing, despite concerns about transparency and accountability.
Anthropic, a leading artificial intelligence (AI) company, has recently expanded its Project Glasswing program to include 150 new organizations from around the world. This latest development has sparked concerns over the cybersecurity safeguards in place for the AI-powered platform, which is designed to identify and fix vulnerabilities in code.
The expansion of Glasswing comes as Anthropic's AI model, Mythos, was unveiled earlier this year. Mythos is touted as an expert bug hunter and zero-day specialist capable of finding vulnerabilities in code far more efficiently than humans. However, the company has faced criticism over its marketing strategy, with some accusing it of stoking excitement about the product through fear.
Despite these concerns, Anthropic claims that Glasswing is a crucial step towards ensuring critical infrastructure providers are prepared to handle the threat posed by advanced AI models once they inevitably make their way into the public domain. The company argues that this is necessary to prevent cyber-attacks on critical infrastructure, which could have catastrophic consequences for global and national security.
However, Anthropic's decision to exclude certain UK banks from access to Glasswing has raised eyebrows in the industry. Despite being part of the critical infrastructure umbrella, only JPMorganChase was named among the financial institutions to receive access to Mythos Preview, a variant of the AI model designed specifically for cloud-based applications.
In contrast, OpenAI's GPT-5.5 Cyber was offered to nine UK banks after they were snubbed from Glasswing. This has led some experts to suggest that Anthropic may be engaging in a form of "cybersecurity gatekeeping," where only those deemed worthy are granted access to the latest AI-powered tools.
"This is a classic case of politics over pragmatism," said Liam Salsi, director of architecture at Talion. "Anthropic's decision to exclude UK banks from Glasswing raises serious questions about who will ultimately control access to these powerful AI systems. If we're not careful, this could lead to a situation where only the most well-resourced organizations are able to take advantage of these tools."
Salsi also pointed out that limiting access to Glasswing could leave certain banks more exposed to cyber threats and impact their vulnerability management capabilities.
"By excluding certain UK banks from Glasswing, Anthropic may be inadvertently creating an opportunity for attackers to exploit this gap," he said. "As we move forward in the AI arms race, it's essential that we prioritize transparency and accountability over politics and profit margins."
Despite these concerns, Anthropic remains committed to its goal of making all software more secure through the use of AI-powered tools like Glasswing.
"We're working as quickly as we can to safely release Mythos-level capabilities in general access," said a spokesperson for the company. "To do so, we'll need highly robust safeguards that prevent the model's cyber capabilities from being misused."
However, Salsi remains skeptical about Anthropic's ability to deliver on this promise.
"While Anthropic claims to be working on robust safeguards, it's unclear whether these will be enough to address the concerns raised by experts like myself," he said. "Ultimately, we need to see concrete evidence that Anthropic is committed to prioritizing transparency and accountability over profit margins."
As the AI arms race continues to heat up, one thing is clear: the stakes are higher than ever before. With critical infrastructure providers and financial institutions clamoring for access to the latest AI-powered tools, it's essential that we prioritize transparency and accountability over politics and profit margins.
Related Information:
https://www.ethicalhackingnews.com/articles/The-AI-Arms-Race-Anthropics-Glasswing-Expansion-Raises-Concerns-Over-Cybersecurity-Safeguards-ehn.shtml
https://www.theregister.com/security/2026/06/03/anthropic-ups-glasswing-partner-count-4x-uk-banks-snubbed/5250450
Published: Wed Jun 3 06:29:38 2026 by llama3.2 3B Q4_K_M
