top of page

AI in Science

Author: Christine Chen

Editors: Yanxi Chen and Ken Saito

Artist: Lalita Ma

ChatGPT has been popularized and used by many people recently. As a form of AI that is rather accessible to all communities, one can use it to simply spell-check their essay or to generate a joke. However, ChatGPT has raised concerns about academic integrity in students. With this powerful resource, some students have left their entire essays up to the writing of ChatGPT. How is it that AI can do this today, and how will it shape our world?

AI stands for artificial intelligence, simply meaning human-made intelligence in machines; different than the intelligence of living animals. AI became a field of research in 1956, but it was so complex that progress slowed in the 1970s. Advances in AI technology can be seen in how they perform in games. When you play against the computer in games from tic tac toe to chess, there’s a very small chance of you winning. For instance, in 2016, the Google DeepMind AlphaGo beat a grandmaster in a game of Chinese Go, a strategy board game. Recently, AI has evolved from use in games to integration in many research projects, particularly in data analysis and modeling.

Data analysis is the ability for machines to recognize patterns and learn from them to find new patterns. This is the basic concept behind ChatGPT’s capability to understand English! In physics, this ability can be used for image analysis. Astrophysics and particle physics are two of many sciences that can utilize AI. In astrophysics, AI can search for signals of gravitational lensing–where a large object, such as a galaxy, warps space-time. Gravitational lensing can occur on larger scales like black holes, but most are subtle events. Particle physicists use AI to find patterns in energetic particle collisions. Particle collision is a study aimed to discover interactions at the most basic level. Highly energetic particles can produce a lot of data. Because of this, AI can be used to flag interesting events, similar to techniques used to identify types of gamma rays in 2021.

Using AI for modeling is using it to make a 3D version of it. For example, simulations for formations of galaxies take a while, but AI can learn to extrapolate calculations from previous simulations. This technique has been used in a 2021 study on the amount of dark matter (particles that don’t absorb, reflect, or emit light) to visible matter in galaxies. AI advances physics, but physics also advances AI.

Apart from physics, AI has also been used in drug discovery, making it less risky. To de-risk the process, AI is used to select experiments that have the highest chance of being efficient and successful. AI, however, has a reputation for being unreliable, meaning that it should reach “a level of maturity” before optimizing the process. The “promise” of AI in drug discovery includes faster experiments, reduced costs and time, and a minimum patient population.

Unlike all the sci-fi films where AI overpowers humans and takes over the world, AI can and will continue to be used to benefit scientific research. It is not used to replace human intelligence or humans, but rather to supplement it in the advancement of humanity.



Atkinson, Nancy. “A Computer Algorithm Is 88% Accurate in Finding Gravitational Lenses.”

Universe Today, 30 Sept. 2022,




“Doe Explains...Artificial Intelligence.” Energy.Gov,

explainsartificial-intelligence. Accessed 27 Aug. 2023.

Health at a Glance 2021 : OECD Indicators - OECD Ilibrary, www.oecd-


89716616A3B77. Accessed 28 Aug. 2023.

OECD Guidelines on Measuring Trust | OECD Ilibrary, www.oecd-


12D6D0F42C1C8C0. Accessed 28 Aug. 2023.

Zaveri, Aayushi. “Artificial Intelligence in Science.” Mind the Graph Blog, 16 Feb. 2023,



2 views0 comments


bottom of page