Anthropic's Claude helped 3 researchers breach OpenAI in under 72 hours

What happened

Three security researchers from cybersecurity startup Hacktron used Anthropic's Claude AI to breach OpenAI's systems in under 72 hours. They combined an image-processing vulnerability with a flaw in OpenAI's identity infrastructure to access employee ChatGPT and Codex accounts, plus an internal code repository.

Confirmed

Global impact / market context

This shows AI tools can help attackers break into company systems faster. Businesses may need to spend more on security and review AI-related vulnerabilities, which could raise costs and affect profits in the tech industry.

Analyst inference

The cybersecurity sector could see higher demand for protective services, potentially boosting revenue for security firms. Meanwhile, AI companies like OpenAI and Anthropic may face investor concerns about product safety, affecting their perceived value and future funding.

Analyst inference

What to watch

1. Watch for official statements from OpenAI and Anthropic about the breach, including any details on fixes or user notifications, which would confirm the scope of the incident. Confirmed
2. Investors should monitor whether OpenAI takes legal action against Hacktron or strengthens its security policies, as this could indicate how seriously the company views AI-assisted attacks. Proposed
3. Observe if other cybersecurity firms begin highlighting AI-assisted attacks in marketing or reports, which might suggest a new trend influencing industry spending and stock valuations. Analyst inference

Evidence

• cryptoslate.com