AI isn’t learning from humans anymore — it’s learning from itself.
Today, Rowan Hale breaks down one of the most important shifts happening inside Google, Meta, TikTok, YouTube, and the big AI labs: a world where AI-generated content is flooding the internet… and then being used to train the next generation of AI.

The result?
A self-reinforcing intelligence loop — where human knowledge becomes downstream of AI.

In this episode, we cover:

• How Google Search is quietly using engagement on AI answers as training data
• How Meta, TikTok, and YouTube boost AI-edited content and train on it
• Why OpenAI, Anthropic, and Google DeepMind are partially training their new models on older model outputs
• How AI-filtered content shapes what humans believe — and then feeds back into future AI training
• The long-term dangers of an intelligence layer that learns from its own reflection

🔗

TikTok’s rising use of AI-generated and AI-edited content
https://www.theverge.com/2023/8/16/23834552/tiktok-ai-label-policy-content-rules

AI-summaries influencing public understanding of news
https://www.poynter.org/fact-checking/2024/how-ai-summarization-is-changing-news-consumption/

Meta boosting AI-generated content across Facebook & Instagram
https://www.engadget.com/meta-is-now-pushing-ai-generated-content-into-your-feed-174549293.html

💡 Key Takeaway

When AI trains on AI-shaped information, the world becomes a feedback loop —not a reflection of reality, but a reflection of AI’s interpretation of reality.

This is the new power center.And most people haven’t noticed it yet.

Hosted by Rowan Hale.