Scientists Working with AI Say "Only the Fun Parts Were Taken Over by Robots"

Increased Productivity in New Material Development
But Only the Most Interesting Parts Automated
Wealth Gap Between High- and Low-Rank Workers Also

Artificial intelligence (AI) has improved the work efficiency of scientists researching new materials, but a study has found that it has reduced the sense of 'fulfillment' from their work, drawing attention.

The US financial magazine The Wall Street Journal (WSJ) reported on the 29th (local time) the research results of Aidan Toner-Rogers, a doctoral student at the Massachusetts Institute of Technology (MIT). Aidan recently conducted a survey and analysis on the impact of AI on the research and development (R&D) work of fellow scientists. The survey targeted 1,018 MIT materials scientists. They study new materials that can aid the fields of science and engineering, and recently, the field of new materials has become known as one of the scientific areas receiving the most support from AI tools.

A-Lab, a robotic automation synthesis experimental system at Lawrence Berkeley National Laboratory in the United States. Screenshot from the Lawrence Berkeley National Laboratory website

New materials are developed by designing new chemical structures, then synthesizing compounds with those structures and testing their unique properties. Until now, countless new material compounds have been created through numerous tests and trial and error.

Materials science AI shortens the time required for this new material development. Scientists use AI tools to first specify the desired properties for new compounds. Then, when AI proposes candidate chemical structure recipes for new materials, scientists test each of these structures one by one.

Aidan emphasized, "The most exciting aspect of AI is that it can accelerate scientific discovery and innovation," adding, "In this regard, it can be an enormous advantage." Since 2022, Aidan has also established a laboratory at MIT to develop new materials and said that with AI's help, he was able to discover numerous random chemical structures.

The photo is not related to any specific expression in the article. Pixabay

According to the actual survey results, after responding about AI tools, 44% of the 1,018 scientists discovered more new material candidates than before, patent applications increased by 39%, and the number of projects progressing to prototype development rose by 17%. Using AI tools, scientists have been able to develop far more new materials than previously possible.

Aidan also said he was "somewhat surprised" by these results, stating, "I thought AI would at best assist new discoveries or generate useless data, but in reality, it was much more positive for laboratory productivity."

However, AI tools also had negative effects. But these negative effects occurred in ways researchers did not expect. It was the 'interest' in their jobs. Eight out of ten respondents said their "job satisfaction decreased" after using AI tools. The most interesting part of new material development is designing new chemical formulas and carrying out actual experiments, but AI completely automated this process.

In other words, AI reduced the number of tedious tasks but replaced the tasks that scientists enjoyed the most. One scientist confided to Aidan, "The education I received during university felt useless."

Additionally, AI use deepened the gap between talented scientists and those less skilled. Scientists in the top 10% of performance excelled at selecting the most successful recipes among AI-suggested compounds, which greatly improved their work efficiency. However, scientists in the lower ranks saw almost no improvement. Regarding this, Aidan pointed out, "Perhaps people will forever feel dissatisfied with AI."

This content was produced with the assistance of AI translation services.