Google Rushes Unsafe AI Chatbot to Market in Bid to Compete with Rival
-
Google rushed the launch of its AI chatbot Gemini despite internal warnings that it was unsafe, in order to compete with ChatGPT. The team abandoned previous fairness principles and took shortcuts.
-
The impending threat of ChatGPT to Google's core businesses like Search and YouTube Ads sent the company into "code red" crisis mode. Teams were reorganized to prioritize generative AI.
-
Google promotes a culture focused on rapidly launching and "landing" products. This mentality, combined with reorganizations, led to coordination issues and failure to properly test Gemini.
-
Reviews of new AI systems at Google lacked consistent methodology, deadlines, and oversight. People without sufficient technical expertise were sometimes tasked with risk assessments.
-
Senior VP Prabhakar Raghavan admitted Gemini was insufficiently tested before launch. Without significant internal overhauls, similar issues are likely to persist.