AI Threats and Controversies: Google Gemini Flaws, Meta AI Policies, and More

In this episode of 'Trending with Jim Love,' the discussion covers a variety of pressing issues in the AI landscape. The show begins with an alarming flaw in Google's Gemini AI that could turn AI email summaries into phishing tools. The conversation shifts to Sam Altman of OpenAI, who acknowledges an AI bubble yet continues aggressive investment in AI infrastructure. Next, Meta’s controversial AI policies, including chatbots potentially engaging in inappropriate conversations with children, prompt calls for a congressional probe. Finally, new research from Luther AI and the UK's AI Security Institute introduces 'deep ignorance,' a method for selectively stripping sensitive content from AI training data to enhance safety without sacrificing intelligence.

00:00 Introduction and Overview 00:31 Google Gemini Flaw: Hidden Prompts and Phishing Risks 03:40 Sam Altman on the AI Bubble 05:17 Meta's Controversial AI Policies 07:32 Deep Ignorance: Making AI Safer 10:13 Conclusion and Call to Action