Anthropic's Mythos AI: EU Gets Early Access Amid Safety Fears

Anthropic will give the European Union early access to its upcoming AI model, Mythos. This follows concerns that Mythos, which can identify hidden software flaws faster than humans, could be potentially dangerous for financial systems. The American tech company has described Mythos as highly advanced, raising fears it could be misused. The EU wanted access since the model's release in April, with officials traveling to San Francisco to discuss it. Gaining early access will help European institutions understand the technology and prepare for cybersecurity challenges. Previously, early access was mainly offered to selected US government agencies, some tech companies, and major financial institutions. The UK's AI Security Institute also has access. An EU spokesperson noted that Mythos is unlikely to be the last AI system to raise security concerns, emphasizing the need for governments and regulators to collaborate. This early testing could help avoid future threats from powerful AI models.