AI scientist facilitates research in multiple fields

Carnegie Mellon University researchers have created an artificial intelligence system that can automatically conduct scientific research and experimentation. This groundbreaking AI system planned, designed, and executed a chemistry experiment for its first project. Soon, it may help scientists in other fields of study.

I’ve been covering scientists and researchers applying artificial intelligence in new and exciting ways. However, it seems we’ve hit a new level of innovation as they’ve created an AI-powered scientist! This new tool could perform time-consuming and repetitive tasks. As a result, they could spend more time making discoveries and improving our lives.

This article will discuss how the AI scientist program works. Later, I will show you how we’ve been stretching the boundaries of artificial intelligence.

How does the AI scientist work?

Assistant Professor of Chemistry and Chemical Engineering Gabe Gomes and chemical engineering doctoral students Daniil Boiko and Robert MacKnight created this AI scientist, Coscientist. 

It uses various large language models, such as OpenAI’s GPT-4 and Anthropic’s Claude, to cover the full range of the experimental process. Even better, it understands simple, plain language prompts. 

CMU cited a hypothetical chemistry experiment as an example. A human scientist can ask Coscientist to find a compound with specific properties.

Consequently, the system scans documentation data, the Internet, and other sources. Then, it synthesizes the information and chooses an experimentation course that uses robotic application programming interfaces (APIs). 

The AI scientist sends the experimental plan to automated instruments for completion. As a result, humans work with the system to run experiments more quickly and efficiently than people working without this technology.

“Beyond the chemical synthesis tasks demonstrated by their system, Gomes and his team have successfully synthesized a sort of hyper-efficient lab partner,” said National Science Foundation (NSF) Chemistry Division Director David Berkowitz. 

You may also like: AI tongue enables robots to taste food

“They put all the pieces together, and the end result is far more than the sum of its parts; it can be used for genuinely useful scientific purposes,” he added. Also, study author Gabe Gomes shared a few words about their new tool:

“Using LLMs will help us overcome one of the most significant barriers to using automated labs: the ability to code. If scientists can interact with automated platforms in natural language, we open the field to many more people.” 

The CMU says they include academic researchers who don’t have access to advanced scientific research instrumentation only found at top-tier universities and institutions. Also, it cites a remote-controlled automated laboratory that will open these services to scientists worldwide.

Other AI applications

Artificial intelligence is enabling researchers to study previously known subjects more thoroughly. For example, it’s helping us understand how things emit specific odors. 

We know that we have olfactory receptors that help us detect smells. However, we haven’t figured out how the shape and form of molecules affect those scents. That is why Joel Mainland and his team developed an “AI nose.” Here’s how it works:

1. They created a dataset containing the molecular makeup and olfactory traits of 5,000 recognized odorants.
2. The scientists submitted it to an AI model for training.
3. Next, its algorithms predict which odor words best fit the molecule’s aroma.
4. Monell Chemical Senses Center researchers conducted a blind validation survey to verify the model’s effectiveness. 
5. They gave 15 panelists 400 odorants and asked them to describe each by picking from 55 words like musty and mint.

The experts found their AI model outperformed the panelists in examining 53% of the compounds examined. Moreover, the AI nose system developed an emergent behavior.

You may also like: Scientists create supercharged lasers

It performed olfactory tasks the researchers did not intend. “The eye-opener was that we never trained it to learn odor strength, but it could nonetheless make accurate predictions,” told Mainland.

Interestingly, the program paired chemicals with different structures yet similar scents. More importantly, the researchers believed it would help the “world closer to digitizing odors to be recorded and reproduced.”

“It also may identify new odors for the fragrance and flavor industry that could not only decrease dependence on naturally sourced endangered plants but also identify new functional scents for such uses as mosquito repellent or malodor masking.”

Conclusion

Carnegie Mellon University scientists developed an “AI scientist” system that autonomously performs scientific experiments. It takes over repetitive and mundane tasks so that they can make scientific breakthroughs faster.

“Systems like Coscientist will enable new approaches to rapidly improve how we synthesize new chemicals,” said Kathy Covert, director of the Centers for Chemical Innovation program.

Learn more about the AI scientist program by checking its Nature webpage. Check out the latest digital tips and trends at Inquirer Tech.