Inside Google DeepMind’s Approach to Ensuring AI Safety

Lila Ibrahim, COO of Google DeepMind, is interviewed about the company’s approach to ensuring AI safety in an increasingly concerned industry. DeepMind‘s ultimate mission is to develop artificial general intelligence, yet Ibrahim believes that new risks, such as bias, safety, and inequality, should be taken particularly seriously. She outlined the company’s four-pronged strategy that involves adhering to the scientific method, developing a multidisciplinary team, producing and publishing principles, and ensuring a diverse talent pool. DeepMind often works with external experts to mitigate ethical risks and invites feedback from a broad range of communities to refine and retrain their models. Their success with AI protein structure-prediction for drug development has been met with a great response from the scientific community. DeepMind‘s responsible approach and advancements in AI demonstrate the benefits of thinking long-term and prioritizing society’s needs over short-term corporate gains.