The public conversation about AI risk has collapsed into a single story about rogue and super-intelligent AI posing an existential threat to humanity. This is not only anxiety-producing, but is crowding out a much more productive conversation about harms that are real and happening right now. In this episode, we untangle three distinct categories of AI risk that keep getting bundled together: autonomous AI as a threat, malicious use of AI by bad actors, and the broader externalities of AI on domains like labour, privacy, copyright, mental health and the information environment. We explore how each category represents a distinct set of concerns, but also how each requires a different policy response, and the vested interests that benefit from keeping the public debate focused on one category over another.
Podden och tillhörande omslagsbild på den här sidan tillhör
ctrl:cyber. Innehållet i podden är skapat av ctrl:cyber och inte av,
eller tillsammans med, Poddtoppen.
