The AI Argument EP46: Grok 3’s Rigged Answers, Ireland’s Copyright Debate, and an Accidental Evil AI

AI copyright laws could be about to change, but should they? A new report from Ireland’s AI Advisory Council recommends giving AI-generated works limited copyright protection while letting creators opt out of AI training. 

Frank thinks that’s a reasonable way to protect artists. Justin thinks it’s a fussy bureaucratic workaround that won’t help Europe keep up in the AI race. Copyright holders, he argues, should have no right to refuse, only the right to get paid. Because AI and robotics will define the next century, and Europe needs to get in the game, not get tangled in red tape.

The debate doesn’t stop there. 

What happens when a team of researchers accidentally trains an AI to be evil? Why did xAI quietly remove Elon Musk and Donald Trump Jr. from Grok 3’s disinformation lists and then blame OpenAI for it? And would the EU let Elon Musk use AI to fire employees?

Chapters

• 0:00 Grok 3’s Rigged Answers, Ireland’s Copyright Debate, and an Accidental Evil AI: The AI Argument EP46
• 1:24 Did xAI rig Grok 3’s answers to protect Musk and Trump?
• 6:31 Would the EU let Elon’s AI fire employees?
• 8:09 Will AI copyright laws protect Ireland’s creatives?
• 24:30 Can OpenAI’s Deep Research really replace experts?
• 29:19 Are Claude 3.7 and GPT-4.5 real breakthroughs?
• 31:44 Did researchers accidentally create an evil AI?