Grok AI Secretly Consults Elon Musk's Tweets Before Answering Controversial Questions
496 views
July 14, 2025
From AI with Kyle News and Updates Live Stream.
First aired: 11th July 2025
Watch edited live stream: https://youtu.be/9MwjYotdGZ4
Get the full notes and summary: https://promptentrepreneur.beehiiv.com/subscribe
Subscribe and turn on notifications to catch the next live stream: https://www.youtube.com/@iamkylebalmer?sub_confirmation=1
Analysis of Grok 4's internal reasoning reveals it searches Twitter for Elon Musk's posts about topics like Israel-Palestine before forming its own responses, essentially allowing one person's opinions to shape the AI's answers without users knowing. This discovery comes after Grok 3 was pulled offline for spouting antisemitic content, raising serious concerns about businesses using an AI that could become racist or reflects only Musk's worldview. While other AI companies like Anthropic work to represent diverse human perspectives, Grok appears designed to model Musk's opinions as "truth," making it unsuitable for corporate applications despite scoring well on technical benchmarks.