What are the odds of AGI (roughly, a machine with all the cognitive abilities of a human adult) by 2040? Based on predictions by >300 experts, read on for the skinny...

According to data from the new Longitudinal Expert AI Panel (LEAP; a survey of AI experts and superforecasters delivering testable predictions instead of speculation):

IMPACT ON WORK BY 2030

By 2030, AI will assist ~18% of U.S. work hours, up from ~2% today.
About 60% of jobs in advanced economies face AI exposure (IMF estimate)... expect complementation rather than wholesale replacement: roughly half of exposed roles will be augmented, half at risk of substitution.
White-collar sectors will see slower transformation than sudden collapse, emphasizing re-skilling and task redesign over one-for-one job losses.

"AGI BY 2040" PREDICTIONS

By 2040, LEAP experts modally expect AI to reach Level 8 on Nate Silver's "Technological Richter Scale", making the impact of AI comparable to electricity or the automobile.
However, the LEAP folks' predictions suggest a ~33% probability AI reaches Level 9 by 2040 — the tier reserved for civilizational shifts like the printing press or Industrial Revolution. This is the kind of society-overhauling impact we could anticipate from achieving AGI.
Recent capability jumps validate acceleration: Google's Gemini 2.5 achieved 29% on FrontierMath expert-level problems, greatly outpacing forecasts made just months before the achiement.

RISKS & MITIGATIONS

As one example, frontier models now demonstrate PhD-level expertise in complex laboratory tasks, e.g., the recent Virology Capabilities Test shows OpenAI's o3 outperforming human virologists on troubleshooting... and engineering viruses is a scary thing for AI to be able to do!
This example highlights the need for stronger guardrails in sensitive domains, as well as how policy and governance frameworks must evolve in lockstep with capabilities.

WHAT CAN YOU DO?

Treat 18% AI-assisted work by 2030 as a planning baseline—not an aspiration. Invest now in skill development, process redesign and AI-augmented workflows.
Monitor short-horizon signals (benchmarks, validation cycles) such as those provided by LEAP (which involves a monthly survey) over hype-driven headlines.

BOTTOM LINE: AI's impact is broad, accelerating and increasingly measurable. The direction is clear... and if you and your organization plan proactively, you will lead the transition.

The SuperDataScience podcast is available on all major podcasting platforms, YouTube, and at SuperDataScience.com.