The 50-Point Gap
Ask an AI expert whether AI will have a positive impact on how people do their jobs, and 73% say yes. Ask a random American, and 23% say yes. That 50-percentage-point gap is the single largest disagreement in Pew's 2024 AI survey — larger than on the economy, medical care, or K–12 education. It's the defining asymmetry of the AI conversation, and it's why AI Takeover Tracker exists.
Key finding: Over a 20-year horizon, 64% of Americans expect AI to result in fewer jobs. Only 5% expect more. AI experts are less pessimistic but still net-negative: 39% fewer, 19% more. Both groups expect reductions — the disagreement is mostly about magnitude, and about who benefits.
Who thinks AI will have a positive impact?
% answering “positive impact”The people who build and study AI are dramatically more optimistic about its workplace impact than the people whose workplaces it touches. The gap is largest on jobs — 50 percentage points between U.S. experts (73%) and U.S. adults (23%).
The long-horizon jobs forecast
64% expect fewer jobs — 12.8× as many as expect more.
39% expect fewer jobs, but 19% expect more — the closest either group gets to balanced.
Both groups forecast a contraction. The public estimates it with near-consensus certainty; experts are split but still net-negative. The actual evidence so far (see our Age Gap and Occupational Churn insights) lands in between: real displacement in specific cohorts, modest change in the aggregate mix.
Who do people trust to regulate AI?
The trust picture is fragmented. Americans report the lowest trust of any surveyed country in their own government to regulate AI responsibly — just 31%. Globally, the EU is the most-trusted regulator of AI, with a median of 53% across 25 Pew-surveyed countries — more than either the U.S. or China. The countries most optimistic about AI (Singapore, Indonesia, Malaysia, Thailand) also show the highest trust in their own governments to manage it.
Why this matters for AI job risk
A 50-point perception gap isn't just statistical. It's the gap between the people making decisions about how AI gets deployed, and the people whose livelihoods depend on those decisions. When a Fortune 500 exec reads “73% of AI experts say AI is positive for jobs,” it shapes capital allocation and headcount planning very differently than if they read “23% of the public agrees.”
The question we try to answer with Takeover Tracker: which group is more right? The honest answer in 2026 is “it depends on who you are.” If you're in a high-capability, low-desire task zone (see our Want vs. Get insight), the public's pessimism tracks the evidence better. If you're in a judgment-heavy, context-rich role, the experts' optimism is closer to right. The task-level scoring in our Capability Coverage Index exists to answer that question at the level of your specific job rather than at the level of aggregate public sentiment.
The other finding worth sitting with: both groups expect net job losses, even the experts. The disagreement isn't about whether AI will displace workers — it's about whether the displacement nets out to something positive for workers. That's a values question, not just a forecasting one.
Sources
- Pew Research Center, 2024 survey of U.S. adults and AI experts — expert-vs-public comparisons on societal impact. AI experts defined as U.S.-based authors or presenters at AI-related conferences in 2023–2024.
- Long-horizon jobs forecast (64% public fewer / 5% public more; 39% expert fewer / 19% expert more over 20 years) from the same Pew survey and the Longitudinal Expert AI Panel (LEAP) conducted by the Forecasting Research Institute.
- Global nervousness (52% in 2025, up from 50% in 2024) from the Ipsos AI Monitor 2025. Trust-in-regulator figures from Pew's 2025 25-country survey.
- Our own Capability Coverage Index methodology is built to be usable regardless of which group's framing you trust — the task-level scoring doesn't require you to pick a side on the expert-public debate.