AI to Outperform Humans Sooner Than Expected: AI Experts

Last Updated:
AI to Outperform Humans Sooner Than Expected: AI Experts
  • AI Impacts survey of 2,700 experts: Accelerated timeline for AI surpassing human capabilities in all activities.
  • High-Level Machine Intelligence (HLMI) is estimated at 10% by 2027 and 50% by 2047, down 13 years from 2022 projections.
  • Full Automation of Labor (FAOL) forecast advances to 10% by 2037 and 50% by 2116, compared to 2164 in the 2022 survey.

In a landscape where technological advancements constantly reshape the employment horizon, a recent AI Impacts survey of over 2,700 AI experts reveals an accelerated timeline for AI systems to outperform humans across all activities. The shift suggests that the era of machines possessing human-level intelligence may be approaching faster than anticipated.

How soon will human-level performance on all tasks or occupations be feasible?

Contrary to previous estimates, the likelihood of High-Level Machine Intelligence (HLMI) hitting 10% by 2027 and 50% by 2047 challenges previous estimates, bringing the latter date forward by 13 years from the 2022 survey. In comparison, the expected date had only shifted one year earlier in the six years between the 2016 and 2022 surveys.

Moreover, the forecast for Full Automation of Labor (FAOL) has accelerated, reaching 10% by 2037 and 50% by 2116, a notable advancement from the 2164 projection in the 2022 survey. Notably, the predictions for a 50% chance of FAOL consistently lag more than sixty years behind those for HLMI.

High-Level Machine Intelligence (HLMI) occurs when machines can outperform humans in every “task” more efficiently and economically. Full Automation of Labor (FAOL) takes place when an “occupation” becomes entirely automatable, with machines accomplishing it better and more cost-effectively than humans.

Despite the accelerated timelines, there is an undercurrent of uncertainty among AI researchers regarding the long-term implications of advanced AI. Approximately 68.3% of respondents expressed optimism, believing that positive outcomes from superhuman AI are more likely than negative ones.

However, within this group of optimists, 48% acknowledged at least a 5% chance of extremely negative outcomes, such as human extinction. Nonetheless, this shows a stark contrast to the 61% of Americans who view AI as a threat, according to an August 2023 poll.

Disclaimer: The information presented in this article is for informational and educational purposes only. The article does not constitute financial advice or advice of any kind. Coin Edition is not responsible for any losses incurred as a result of the utilization of content, products, or services mentioned. Readers are advised to exercise caution before taking any action related to the company.