AI Drives Faster Discoveries But Lowers Job Satisfaction, Study Finds

Artificial Intelligence is making a big impact on materials science, speeding up discoveries and leading to more patents and new products. A recent study looked at how an AI tool for materials discovery affected a major research lab, showing both exciting benefits and significant challenges.

The study found that AI can greatly increase the rate of materials discovery by automating many idea-generation tasks. This allows scientists to focus on testing the suggestions made by the AI. Analyzing the mechanisms, the researcher found that AI automates 57% of idea-generation tasks, shifting researchers to focus on evaluating AI-generated candidate materials.

However, the success of this technology relies heavily on the skills of the researchers using it. Scientists with strong judgment and expertise quickly learned to prioritize the AI’s best suggestions, while others struggled, wasting time on ideas that didn’t work.

Despite these benefits, the study revealed a major downside: most researchers felt less satisfied with their work. About 82% reported lower job satisfaction, mainly because their tasks became less creative and more repetitive. Many said they felt their skills were underused, and the shift in focus made the work less enjoyable.

These findings challenge the idea that AI will only take over boring tasks, leaving people to focus on more rewarding work. Instead, the study showed that AI is automating some of the most exciting parts of scientific research—like coming up with new ideas.

“I couldn’t help feeling that much of my education is now worthless,” One scientist from the study said, as reported by The Wall Street Journal.

The study also looked at how the lab adapted to the changes AI brought. After seeing how the tool worked, the lab changed its hiring practices to focus on scientists with strong judgment skills. While this helped improve collaboration with the AI, it widened the gap between more and less productive researchers, adding to the dissatisfaction for some.

The author believes AI could help reduce income inequality if workers receive proper training. For instance, a well-trained paralegal using AI could handle tasks typically done by experienced lawyers, potentially earning a higher income.

Without adequate training, however, the situation is “like sending people up in an airplane without putting them in a flight simulator first and telling them, ‘Oh, sorry, you crashed, I guess you’re a bad pilot,’” he explained to The Wall Street Journal.

The findings raise important questions about how AI will affect the future of scientific work. While AI tools can accelerate innovation, they also risk making jobs less engaging if not used carefully.

The study suggests that organizations and policymakers need to find ways to balance these effects—such as offering better training and redesigning roles to keep tasks interesting and rewarding.