Researchers are testing a new and advanced artificial intelligence, or AI, model to see how it handles hard tasks in areas such as in mathematics and neuroscience. Released by a Chinese company, DeepSeek-RI is an inexpensive and powerful tool being described as a technological game changer in enabling researchers to advance their work. But it’s also been found to fail at some tasks researchers might want to perform. As usual with Chinese AI models, experts note that it won’t answer politically sensitive questions. But AI expert Subbarao Kambhampati, a professor in the School of Computing and Augmented Intelligence, part of the Fulton Schools, says that censorship looks more like a clumsy add-on than something baked into system.