You probably picture a bunch of computer servers when thinking of generative AI labs. Well Anthropic is swapping some of that for actual test tubes and petri dishes. The creators of Claude recently admitted they have a wet lab located in the San Francisco Bay area. So instead of just using their AIs to simulate biological functions they can now try them in a real life lab setting.
There’s only so much you can learn from simulating things in the field of biology. “In the end there will always need to be a physical lab to prove your hypothesis” Anthropic’s head of life sciences Eric Kauderer-Abrams stated. You can build amazing theories through the use of AIs but in order to test those theories on real life organisms you’ll need a lab. They plan to do some lab work themselves as well as partner with other labs. Which gives them a similar setup to that of most biotech companies. Using lots of algorithms along with physical testing in a lab.
Of course people assume they want to get into the pharma industry when you hear that they’re opening up a lab. They state that their lab is more focused on basic science and biology. Which makes sense since they already have many large pharma companies using their services. If they were to start working on making drugs it may deter those companies from using their platform.
They’ve clearly made a large investment into their life science sector. After buying the biotech company Coefficient Bio earlier this year they’ve been investing heavily into their life sciences team. Whether it be number of employees or amount of money spent they’ve grown rapidly. They’ve also started experimenting with automated lab equipment and if they can allow their model to run it with limited human interaction.
It’s interesting that Anthropic wants to delve further into biology when they’ve warned about many dangers of AI. They have previously brought attention to the dangers of AI developed bio weapons. So its kind of ironic that they want to use AI in a lab to discover new information about life. But they claim that by gating their model and keeping it safe they believe it can be done. If they can successfully use their AI to prove a theory in the lab it could open many doors in the scientific community.