Home  

My Websites 

My Bio

 

The document “Thousands of AI Authors on the Future of AI” presents the findings of a survey of 2,778 researchers who had published in top-tier artificial intelligence (AI) venues. The survey asked for their predictions on the pace of AI progress and the nature and impacts of advanced AI systems. The aggregate forecasts give at least a 50% chance of AI systems achieving several milestones by 2028, including autonomously constructing a payment processing site from scratch, creating a song indistinguishable from a new song by a popular musician, and autonomously downloading and fine-tuning a large language model. The survey also found that if science continues undisrupted, the chance of unaided machines outperforming humans in every possible task was estimated at 10% by 2027, and 50% by 2047. The survey revealed substantial uncertainty about the long-term value of AI progress, with most respondents expressing uncertainty about the outcomes of superhuman AI, and a significant percentage giving at least a 10% chance to advanced AI leading to outcomes as bad as human extinction. The survey also highlighted disagreement about whether faster or slower AI progress would be better for the future of humanity, but there was broad agreement that research aimed at minimizing potential risks from AI systems ought to be prioritized more.

The revision among the experts cited was significant. The aggregate predictions for 21 out of 32 tasks moved earlier between 2022 and 2023, while the aggregate predictions for 11 tasks moved later. On average, for the 32 tasks included in both the 2022 and 2023 surveys, the 50th percentile year they were expected to become feasible shifted 1.0 years earlier. For example, the aggregate 2023 forecast predicted a 50% chance of High-Level Machine Intelligence (HLMI) by 2047, down thirteen years from 2060 in the 2022 survey. There was also a notable shift in the framing effect of fixed-years versus fixed-probabilities, with substantial disagreement among participants about the likelihood of an ‘intelligence explosion’ and the traits that AI systems in 2043 would possess. Additionally, respondents expressed significant concern about potential AI scenarios, with more than 30% considering each scenario worthy of substantial or extreme concern[1].

So what are we to make of this? Well, for one, AI is progressing faster than anyone has anticipated. Two, we all need to revise our future focus, and start preparing for many of the changes that were predicted to be decades away, that now most likely are just months to years away.

Citations:

https://youtu.be/QdCJ3YOVVtc?si=p38fZ-7Zneq-jtlv

[1] https://aiimpacts.org/wp-content/uploads/2023/04/Thousands_of_AI_authors_on_the_future_of_AI.pdf

©  All Rights Reserved, Intramation Technology