In 2017, John Jumper, who had recently completed a PhD in theoretical chemistry, applied for a position at Google DeepMind after hearing about a project aimed at predicting protein structures. By 2020, he, along with CEO Demis Hassabis, developed AlphaFold 2, an AI system that accurately predicts protein structures to an atomic level, achieving results significantly faster than traditional laboratory methods.
In 2021, Jumper and Hassabis were awarded the Nobel Prize in Chemistry for their work. As the excitement surrounding AlphaFold begins to settle, questions arise about its real-world implications and the ways in which researchers are utilizing the technology. Jumper discussed these topics recently, alongside insights from other scientists.
Meanwhile, the landscape of AI is evolving, with recent studies indicating that generative AI primarily serves companionship functions. Platforms such as Character.AI and Replika allow users to create personalized chatbots that can embody various personas, including friends or therapists. This trend has caught the attention of some state governments, which are beginning to regulate companion AI. However, a significant gap remains in addressing user privacy within these regulations.
This information is part of the collaboration between the Financial Times and MIT Technology Review, titled “The State of AI.” Subscribers to their respective newsletters can access additional content on these developments in AI.
Source: https://www.technologyreview.com/2025/11/25/1128346/the-download-the-future-of-alphafold-and-chatbot-privacy-concerns/