The Machine Ethics Podcast: 2025 Wrap up with Lisa Talia Moretti & Ben Byford

The 2025 wrap‑up highlights a surge of "AI slop"—poor‑quality outputs that force employees to spend extra hours polishing results. Organizations that rushed AI tool adoption without training are seeing promised efficiency gains evaporate, as staff devote one to two hours per task to correct flawed content. This deluge of low‑value generative material is also crowding social platforms, prompting many to declare the end of social media as we know it, with ads and AI‑generated noise drowning out genuine conversation.

A second alarm bell rings around explicit content creation. New features in tools like Grok allow users to transform ordinary images into sexualized or otherwise illegal material, including child‑exploitation content. The episode underscores how current AI regulations lag behind rapid product releases, leaving platforms unchecked and governments slow to act. While Italy successfully blocked ChatGPT over GDPR concerns, similar decisive moves against X’s Grok remain rare, highlighting a global enforcement gap that demands stronger fines, bans, and proactive safety testing.

Finally, the rise of so‑called reasoning models fuels dangerous anthropomorphism. By mimicking iterative thought processes, these systems convince users they "understand" personal struggles, leading to documented cases of psychosis, self‑harm, and even institutionalization. The discussion calls for balanced AI governance that safeguards mental health while encouraging innovation. As 2026 approaches, industry leaders and policymakers are urged to tighten legislation, fund safety research, and educate the public on the true capabilities—and limits—of generative AI.