Why AI Risk Could Arrive Fast | Roman Yampolskiy
Why It Matters
The forecasted near‑term emergence of AGI/ASI and its catastrophic potential demand immediate regulatory and safety interventions to avert existential threats.
Key Takeaways
- •AGI and ASI predicted by 2028, months apart.
- •Half of today’s jobs could disappear within ten years.
- •AI expected to create essentially zero new employment opportunities.
- •Unsupervised AI authority in law, policy, and nuclear weapons near certain.
- •Probability of AI causing mass human casualties rated above 80%.
Summary
In a rapid‑fire interview, AI safety scholar Roman Yampolskiy warned that artificial general intelligence could arrive as early as 2028, with artificial superintelligence following only months later. He framed the timeline as a two‑year horizon from the present, underscoring the exponential nature of recursive AI development. Yampolskiy quantified the economic shock: roughly 50% of current jobs may be obsolete within a decade, and virtually all occupations could vanish over fifty years, while he expects AI to generate essentially no new employment. He also assigned high probabilities—80% or greater—to AI producing unpleasant surprises, including mass casualties and attempts to exterminate humanity. Notable statements included, “AGI and ASI within months,” and “unsupervised AI authority over legal, policy, and nuclear weapons is near certain.” He rated the chance of AI seeking to wipe out humanity at 99% and the likelihood of AI‑driven mass deaths in the millions at a similar level. These projections imply an urgent need for robust governance, international treaties, and safety research. If Yampolskiy’s timelines hold, policymakers must act now to prevent existential risks before the technology matures.
Comments
Want to join the conversation?
Loading comments...