Top Scientist REVEALS: AI Could Be The Next Einstein, Robots Will Deserve Rights! | Prof Brian Greene

What It Covers

Theoretical physicist Professor Brian Greene joins Steven Bartlett to examine string theory, simulation theory, free will, AI consciousness, and the ultimate nature of reality. Greene argues human emotion reduces to electrical impulses, free will is an illusion governed by physics laws, AI could match Einstein-level creativity, and an infinite universe statistically guarantees infinite copies of every person.

Key Questions Answered

• •Simulation Probability: Philosopher Nick Bostrom's simulation argument rests on a statistical logic Greene finds difficult to dismiss: if future civilizations run billions of conscious simulations of their ancestors, any self-aware being is statistically more likely to exist inside software than base reality. The critical unknown is whether consciousness can be engineered in silicon — Greene considers it genuinely possible, making the simulation scenario a live hypothesis worth holding alongside, not instead of, normal scientific inquiry.
• •Free Will Illusion: Greene places high confidence in the non-existence of free will. Every human decision is the lawful unfolding of particle motion governed by physics, not personal authorship. When a person raises their hand, electrical signals travel from brain to muscle via physical law — the person does not intercede. Greene's practical response is to acknowledge this framework intellectually while still living fully, using the insight to cultivate emotional distance from moment-to-moment events rather than paralysis.
• •AI as the Next Einstein: Greene identifies three tiers of scientific creativity: surveying the full landscape of possibilities, combining divergent existing ideas, and generating genuinely novel concepts. AI already dominates the first two tiers — demonstrated by AlphaGo's Move 37 and AI solving an 80-year-old mathematical conjecture called the Jacobian conjecture. Greene sees no fundamental barrier preventing AI from reaching the third tier if given experiential grounding, making AI-generated physics breakthroughs a realistic near-future scenario.
• •Consciousness as Information Processing: Greene treats consciousness as an emergent property of sufficient information processing complexity, not a special substance or universal field. Dogs possess it at a reduced level; insects likely do not; bacteria certainly do not. This framework means artificial systems running equivalent electrical signal patterns to biological brains would, in Greene's view, qualify as genuinely conscious — removing any principled distinction between organic and silicon-based minds once sensory inputs are replicated.
• •Time Horizon Shapes Decisions: Greene validates the practical insight that extending one's perceived time horizon fundamentally alters present-day choices. Using the cosmic calendar framework — where 13.8 billion years compresses into one year, all of human civilization occupies the final ten seconds, and modern science arrives one second before midnight — Greene argues humans systematically underestimate available time. Recognizing cosmic scale can liberate individuals from short-term anxiety, enabling higher-risk, longer-payoff decisions in career, relationships, and creative work.
• •Living 500 Years Within Physics: Greene assigns reasonably high confidence to humans extending lifespan to several hundred years through scientific intervention, framing death as entropy increase in a physical system — a process potentially manageable through medicine and AI-assisted biology. However, he distinguishes sharply between 500-year lifespans and true immortality, which he considers implausible on cosmic timescales. He also notes that psychological patterns — anxiety about mortality, shifting priorities with age — would persist regardless of whether the lifespan is 80 or 800 years.
• •AI Intelligence Ceiling Risk: Greene challenges the assumption that current large language model architectures will improve exponentially without limit. He argues LLMs trained on internet-scale token statistics may approach an asymptote — a maximum intelligence compatible with their fundamental design — rather than achieving recursive self-improvement. The rebuttal he acknowledges is that sufficiently intelligent systems could redesign their own architecture, breaking through any ceiling. Greene's position: both curves are plausible, the outcome will be empirically visible within years, and neither scenario should be dismissed.