I recently had the pleasure of recording three engaging conversations with Dr. Nikil Mukerji – covering Deepfake detection and my upcoming SkepKon talk on the topic, critical discourse analysis and media critique, as well as the background and implications of the conflict within the German skeptics’ organization GWUP.
This post includes all three videos – plus a cartoon to celebrate the latest orthographic superpowers of AI generators.
Note: All videos are in German, but you can turn on auto-translated English subtitles via YouTube’s settings.
Media Literacy in the Age of Generative AI (5 min)
How will Deepfakes and AI shape our future? Can we still tell what’s real and what isn’t – and if so, how?
This short video offers an introduction to the capabilities of generative AI and the latest methods for detecting Deepfakes. For a deeper dive, join me at SkepKon 2025 on May 30, where I’ll be exploring the topic in detail.
Cognitive Dissonance, Discourse Analysis & Media Critique (29 Min.)
“If you feel cognitive dissonance, lean into it!” – That’s one of the takeaways from this discussion, where we talk about the methods of critical discourse analysis, media spin, and the pursuit of objectivity.
Skepticism in the Age of Filter Bubbles (>1h)
In this extended conversation, Nikil and I explore the internal conflict currently unfolding within GWUP – a topic I previously covered in Skeptical Inquirer. What sparked the controversy, and how did it escalate so dramatically? We investigate the causes on both a personal and societal level.
AI Can Spell Now – and That’s Just the Start
It’s not just Ghibli-style illustrations anymore: in recent weeks, DALL·E and other models have learned to integrate complex typography into generated images. Whether it’s fake movie posters, action figures, or signage in urban scenes – the results are surprisingly accurate. The days of spotting AI-generated images by their nonsense text are definitely over!
To celebrate this milestone – here’s a little cartoon:
Oh, and OpenAI’s new o3 model just scored ~132 on the Mensa IQ test – over 30 points more than the best models did 12 months ago. With progress toward the singularity accelerating this rapidly, my talk on Deepfake detection for SkepKon is under constant revision. A curated resource list featuring the best current detectors, compelling research papers, and helpful tools will follow in the coming weeks – so check back soon!
