Key facts
- European lawmakers criticized Anthropic for sending a junior technical employee, Donny Greenberg, to testify on AI safety risks.
- Greenberg, who joined Anthropic in April, testified remotely before the European Parliament.
- Lawmakers had requested Anthropic's head of public policy, Sarah Heck, to attend the hearing.
- The company's decision led to unanswered policy questions and criticism that Anthropic disregards European concerns.
- Greenberg stated he was a technical, not a policy, person and that Anthropic is collaborating with EU bodies on AI safety.
European policymakers expressed dissatisfaction with AI company Anthropic after it sent a recently hired technical employee, Donny Greenberg, to testify remotely before the European Parliament regarding AI safety risks. Lawmakers had specifically requested the presence of Anthropic's head of public policy, Sarah Heck, but the company instead assigned Greenberg, who joined in April following Anthropic's acquisition of his company, Runhouse.
During the hearing, Greenberg emphasized his technical background, stating, "I'm a technical person, not a policy person," which led to frustration among parliamentarians. Dutch Greens lawmaker Kim van Sparrentak noted that Anthropic could have anticipated policy-related questions and suggested the company's actions indicated a lack of care for Europe. Anna Cavazzini, chair of the Parliament's internal market committee, warned Anthropic of future exchanges, while Spanish conservative lawmaker Pablo Arias Echeverría questioned if Greenberg was reading AI-generated responses.
Greenberg defended Anthropic's efforts, stating the company is working with the Commission's AI Office and ENISA to enhance cyber resilience and that AI models are dual-use tools. The hearing concluded awkwardly with Greenberg no longer visible on screen. The incident occurs as the question of how advanced AI models are tested becomes a critical geopolitical and business issue, with rivals like OpenAI and Google DeepMind also engaging with European regulators on AI safety and development standards.