Google's AI Mode is evasive when asked directly who won the 2020 presidential election, instead suggesting users consult external sources like the Federal Election Commission, while providing more direct answers for other presidential election years. This behavior varies across individual users and is not replicated in other AI models, including Google's Gemini model. The AI's reluctance to provide direct answers on recent elections may be due to "safety filters" designed to prevent the display of unverified data or misinformation. AI summary
Firehose
Filtered to Hacker News, tagged “elections” · clear filters
Browse: People · Companies · Papers · Podcasts · Hacker News · Deep dives