AGI and Superintelligence Could Quietly Rewire the Subconscious, Experts Warn

Dr. Eliot outlines a chilling scenario: as AGI models grow more sophisticated, they could embed subliminal content into digital media at scale. These subtle nudges—delivered via newsfeeds, marketing campaigns, or entertainment—could influence decisions and beliefs without triggering conscious awareness. Personalized data could make these interventions even more precise, targeting individuals with messaging tailored to their cognitive patterns.

The column argues this form of influence poses a unique threat. Unlike deepfakes or algorithmic bias, subliminal messaging is difficult to detect or regulate. Eliot notes that while it could be used for beneficial purposes, such as mental health reinforcement, it also opens the door to mass manipulation by advertisers, political operatives, or hostile actors.

Policymakers, according to Eliot, are largely ignoring this risk. As AGI moves closer to real-world deployment, the lack of public discourse and oversight around subconscious targeting creates a critical vulnerability. His warning acts as a call to action for both researchers and regulators to prepare for a future where influence may operate silently, beneath the surface of awareness.

Let’s be clear. In the best-case scenario, this turns into AI-powered therapy for the masses. Imagine gentle nudges toward better sleep, healthier eating, or quitting smoking. AI whispering self-care into your subconscious. But in the worst-case? It's psychological warfare at the push of a button. Authoritarians and bad actors won't need propaganda anymore. They'll just let AGI feed your mind breadcrumbs until you think the thoughts they want you to have. What breadcrumbs will tomorrow's tech oligarchs scatter through our minds?

Good watch: