Ginkgo Bioworks Moves Lab Work From Humans to Robots and AI
Ginkgo Bioworks, a Boston-based biotechnology company, now uses robots and artificial intelligence to run experiments that scientists once performed by hand. The autonomous lab houses one-armed robots enclosed in glass, each working on separate projects while a screen displays the day's tasks and an oversized toy train-like track delivers equipment between stations.
The shift represents a fundamental change in how science gets done. Reshma Shetty, co-founder and chief operating officer, said the turning point came when she first read a lab notebook entry written by an AI model. "The really, really wild moment was the first time I saw a lab notebook entry written by the model," she said.
Four MIT graduate students founded the company nearly two decades ago on a simple premise: programming cells would matter more than programming computers. Jason Kelly, another co-founder, recalls the early skepticism. "We were living on ramen, buying equipment on eBay, and we could not raise venture capital," he said.
The company's fortunes changed after the AI boom accelerated. Kelly read a 2014 blog post by Sam Altman, who later founded OpenAI, about automating biotechnology. The two started talking, and Silicon Valley funding followed.
AI Designs Experiments, Robots Execute Them
Ginkgo's work spans pharmaceutical, agricultural, and government contracts. Current projects include engineering microbes for better fertilizer and creating proteins that produce snow or ice.
Scientists translate experimental designs into instructions for robots. Recently, the company pushed further by having AI design the experiments themselves. Working with OpenAI's ChatGPT, Ginkgo asked the model to create a specific protein - a task typically reserved for human scientists.
The results exceeded expectations. The AI-designed protein synthesis cost 40 percent less than human-designed work and ran more than 30,000 experiments in six months. The company published the results, though the paper has not undergone peer review.
Shetty said the shift has changed her practice fundamentally. "Normally, I rush through designing my experiment because I need to get it done so that I can actually do all the pipetting in the lab," she said. "Now, I spend more time designing my experiments so that the robot can do them for her overnight."
Both Shetty and Kelly stress that humans remain essential. Scientists must still pose the right questions and set appropriate constraints for experiments.
Biosecurity Concerns Emerge
The same accessibility that benefits researchers poses risks. Drew Endy, a bioengineering researcher at Stanford, warns that AI could enable people with minimal scientific training to run experiments with questionable aims.
Endy and colleagues recently published research showing how AI could be used to mass-produce viruses or create biosecurity threats. "I'm thrilled about AI and science right now as a researcher," Endy said, "but I'm not excited about" potential bioweapons programs in other countries.
Historically, biotechnology has been insulated from these risks through intellectual gatekeeping. "Biology has traditionally been hard for people to really gain control over," Endy said. "AI could nudge it a little bit more towards concentration of power."
Regulations and policy to mitigate these risks are achievable but need prioritization before a biotechnological disaster occurs, Endy said.
Kelly acknowledges the coming shift. "I do think you'll have a culture clash coming of what happens when everyday people can ask scientific questions," he said.
Learn more: AI for Science & Research or explore Generative AI and LLM applications in research.
Your membership also unlocks:
