Date:
September 12, 2024
Chai, an AI startup backed by OpenAI has introduced its latest AI model, Chai-1, which will play an essential role in empowering drug discovery.
Artificial Intelligence is rapidly evolving to enable humans build and delpoy better healthcare methologies. OpenAI has been an active investor in innovative AI startups focused on improving healthcare. One such AI startup is Chai, a drug discovery team founded by OpenAI and Meta researchers. The startup has recently raised $30 million to develop drug discovery AI models.
Chai has introduced its first public AI model, Chai-1, to help the medical and life sciences community fasten the drug discovery process and conduct experiments at drastically lower costs. The company has made its AI model’s code and tools available to worldwide researchers without any cost. The only catch is that the tool should not be used for commercial purposes.
The AI model was first made open-source on September 9, 2024 through a free web interface to prompt more research and development in the field. Chai-1 is capable of predicting biochemical molecule structures to speed up drug development. According to its research team, Chai-1 outperforms Google’s Deepmind startup’s AI model AlphaFold in multiple benchmarks.
The AI tool is equipped with the tech to work on various molecules, proteins, DNA, RNA, and even on chemical modifications. One unique feature of the AI is that it can run accurate tests with limited data, something that was not possible till now. It will help conduct experiments using inputs like raw molecular data, experimental findings and sequence information. If the accuracy becomes consistent, the AI model can essentially eliminate the need for Multiple Sequence Alignments (MSAs) to achieve accurate results.
As of now, AlphaFold is the most advanced AI model for advanced protein prediction. But, the outperforming factor of Chai-1 sparks the potential to change this picture and fasten the process of drug discovery at early stage with limited data.
