Anthropic CEO Dario Amodei defended the company’s approach to AI safety, calling it a “race to the top” in a lengthy social media post. He was responding to an AI safety advocate who accused Anthropic of prioritizing the race to superintelligence over its safety commitments. Amodei argued that Anthropic can better influence industry safety standards by remaining a leading AI developer.
Amodei listed nine instances in which he said Anthropic incurred “substantial costs” to prioritize safety. Examples included restricting access to its most powerful models and limiting sales to certain foreign entities.
He also acknowledged missteps, saying the company “could have handled better” a controversial draft of a federal AI bill. Critics argued the proposal would make it harder for the government to regulate dangerous AI models.