AI research is leading to AI that can conduct research on its own. That’s the dream of Japanese company Sakana AI, which has unveiled what it says is the first “AI scientist.” The integrated generative AI model is designed to conduct scientific research autonomously and could lead to discoveries and innovations faster than would be possible with human researchers alone.
Sakana worked with the University of Oxford and the University of British Columbia to design the AI Scientist and published a research paper on what they had done so far. The AI Scientist is supposedly capable of emulating the entire process of scientific discovery, from generating ideas to designing and running experiments and even writing full scientific papers detailing its findings for publication. Sakana said it would also be a bargain, as the model can produce a full research paper for as little as $15.
To get started, all you need is a broad topic and access to data. The AI can then review past research and look for unanswered questions and ways to produce experiments based on those ideas. It can then run in a loop and refine its ideas based on the results of its past research.
For now, the AI scientist is focusing on machine learning – a wise choice, as it is the seed of his own creation. So far, the AI scientist has come up with previously unexplored techniques for different language models and ideas for analyzing how a model evolves from training. That could lead to better machine learning algorithms, and in a way, to successors to the AI scientist.
“The AI scientist automates the entire research lifecycle, from generating new research ideas, writing the necessary code, and running experiments, to summarizing experimental results, visualizing them, and presenting the findings in a full-fledged scientific manuscript,” Sakana explained in a blog post.
“We believe this work marks the beginning of a new era in scientific discovery – bringing the transformative benefits of AI agents to the entire research process, including AI itself. The AI scientist brings us closer to a world where unlimited and affordable creativity and innovation can be unleashed to solve the world’s most challenging problems.”
AI Science
In theory, the AI scientist could branch out into almost any field of research. Environmental protection, medical research, materials science, and more could all benefit from the model’s efforts. However, such a revolution in science is a long way off. AI software is almost at the limit of AI’s expertise, and expanding its capabilities to other scientific disciplines would require much more work. Like any other AI model, hallucinations could arise, meaning the research would be incorrect or downright absurd and worthless to scientific research.
Sakana isn’t the only one who thinks AI could one day conduct its own research. In fact, there may soon be a lot of talk about it. The frenzied rumors surrounding OpenAI’s pending “Strawberry” project suggest that the creator of ChatGPT will have its own AI models capable of conducting research on its own. Admittedly, Strawberry’s idea sounds more like AI doing online follow-up research and thinking ahead to anticipate human questions, but it’s easy to see how that could be adapted to scientific research.
