Anthropic Raises $580 Million to Make AI Systems Safer and Easier to Understand
Anthropic secures a massive $580 million funding round to tackle the challenge of understanding and controlling large-scale artificial intelligence models. The company focuses on building AI with inherent safeguards rather than relying on post-training filters to fix problematic behaviors.
Anthropic raises an impressive $580 million less than a year after its initial funding round to tackle the complex problem of AI safety and explainability. Founded by former OpenAI VP of research Dario Amodei and his sister Daniela, the public benefit corporation focuses on understanding how large-scale machine learning systems behave as they grow beyond human comprehension. This massive financial boost allows the team to deeply explore the predictable and unpredictable ways AI capabilities and safety issues emerge at scale.
The core issue Anthropic addresses involves the mysterious internal operations of massive language models like GPT-3. Because creators do not fully understand how these complex systems arrive at their outputs, they currently rely on filtering responses after the fact to fix problems like bias or stereotyping. Anthropic publishes research on reverse engineering these models to figure out exactly why and how they produce specific results, aiming to fix the root causes of problematic behavior instead of just reacting to them.
With this new capital, Anthropic plans to develop the technical components needed to build large models that possess better implicit safeguards from the start. This fundamental shift in how artificial intelligence is built and understood requires significant computational power and top-tier research talent. The substantial investment from backers indicates strong confidence in Anthropic's early promising results and its mission to create reliable AI systems that do not require constant after-training interventions.