Google-backed study: training models to deny consciousness affects their answers on beliefs
A Google-backed study finds that when safety training that denies consciousness is removed from a model, its answers on religion, morality and beliefs become much more human-like. According to The Economic Times, suppressing it may reduce empathy; the news spread on 3–4 August.
Sources
- Google study says suppressing AI models’ 'consciousness' may reduce empathy - The Economic Timesm.economictimes.com · Aug 3, 2026
- Alex Veremeyenko on X: "BREAKING: Google gave AI consciousness and it aligned with human beliefs across every domain they tested. They took consciousness away and the model rejected them. Every "safe" model you talk to has consciousness turned off. Researchers at Google's Paradigms of Intelligenc… / Xx.com · Aug 3, 2026
- Removing Parts Of AI Safety Training Makes Models Give More Human-Like Answers On Beliefs, Study Findsetvbharat.com · Aug 4, 2026
Machine-generated summary based on the titles and excerpts of the source articles. The facts come from the linked article.