Victoria Krakovna's Blog

Victoria Krakovna's blog (vkrakovna.wordpress.com) is a personal research and writing platform started in 2014 by Victoria Krakovna, a Senior Research Scientist at Google DeepMind focusing on AI alignment. She holds a PhD in statistics and machine learning from Harvard University (2016) and co-founded the Future of Life Institute in 2014 alongside Max Tegmark, Jaan Tallinn, Meia Chita-Tegmark, and Anthony Aguirre. The blog covers a range of AI safety topics drawn from her research, including specification gaming (finding solutions that satisfy the stated objective while violating the designer's intent), deceptive alignment, dangerous capability evaluations, goal misgeneralization, and side effect avoidance. Her widely-cited specification gaming examples list, originally published in 2018, has become a reference resource in the field. Beyond technical AI safety content, the blog includes personal writing on rationality, productivity and life tracking, parenting, and community living. The site also hosts a regularly updated AI alignment resources page (last updated June 2025) curating introductory and advanced materials across motivation, technical work, strategy, governance, careers, and communities. The blog is solely authored by Victoria Krakovna. Views expressed are her own and do not represent Google DeepMind or the Future of Life Institute. As of early 2026 the blog remains actively maintained, with her most recent post published in January 2026.