Anthropic has established the Anthropic Institute, a research unit focused on studying the societal effects of AI and informing policy responses to risks from more advanced systems.
“In the five years since Anthropic began, AI progress has moved incredibly quickly. It took us two years to release our first commercial model, and just three more to develop models that can discover severe cybersecurity vulnerabilities, take on a wide range of real work, and even begin to accelerate the pace of AI development itself,” the company said in the announcement.
Anthropic predicts far more dramatic progress will follow in the next two years.
The company believes rapid advances in AI will force governments and industries to confront difficult questions about jobs, economic disruption and system governance. It also raises concerns about how AI systems express values, how those standards are set and how future self-improving systems should be monitored and regulated.
The goal of the institute is to share what researchers are learning about these challenges as frontier AI systems are developed and to work with external partners to address emerging risks.
The institute will be led by experts and supported by a team of machine learning engineers, economists and social scientists at Anthropic. It builds on several existing research groups focused on AI system testing, real-world deployment and economic impact, with a strong emphasis on incubating new teams.
“The institute has a unique vantage point, with access to information available only to builders of frontier AI systems. It will use this position to report candidly on what is being learned about the development of the technology,” the company noted.
It will also engage workers facing displacement, affected industries and communities struggling to respond to technological change.
Anthropic has come under public scrutiny in recent weeks after the U.S. government suspended the use of its AI tools in government institutions. The company has responded with a lawsuit challenging its designation as a supply chain risk.
