Brandon Zhang :)

Despite the disadvantages and misuse of AI (such as AI Brainrot…) in the daily world, AI has generally been very beneficial to scientists. In this post i will discuss how it does.

AI is changing chemistry by increasing the productivity of scientists which subsequently allows them to make better guesses. In addition, it reduces wasted effort in the lab. Rather than testing every possible molecule or material one by one, researchers can use AI to sort through huge amounts of data and identify the most promising molecules first which saves huge amounts of time. As a result, this makes early-stage chemistry research significantly more efficient because it saves time and helps scientists focus on the options that are most likely to succeed.

An example of this faster discovery is Google’s GNoME using Graph Neural Networks to discover 380,000 new crystal structures. This amount is more than the total known in human history up to right now.

AI is so advanced in the chemical field that AI-generated catalysts are already outperforming traditional catalysts. Some examples include, Hydrogen fuel cell catalysts and CO2 reduction catalysts. unfortunately, it isn’t perfect. While a lot of these designs would theoretically be faster and more productive than current designs, it’s cost and complexity poses a challenge for it to be implemented in real life.

Another barrier is AI “black boxes”. This is when AI reaches a conclusion that seems remarkable but lacks the steps to actually reaching the conclusion. This lack of interpretation reduces scientific understanding and trust

All in all, AI, although flawed in many ways, is still a very powerful tool in chemistry that chemists hope can assist them more in the future