Timeline For AI's Ascendance
Researcher Katja Grace at the University of Oxford's Future of Humanity Institute and a team surveyed 1,634 of the leading artificial intelligence researchers from all over the world about when they believe intelligent machines and the AI that powers them will surpass human intelligence in a variety of contexts. 352 of the experts responded, and the team then calculated median responses. The results of the probe were presented this month.
The experts predicted that within the next decade, AI will outperform humans in tasks like driving trucks (by 2027), translating languages (by 2024), and writing high school essays (by 2026). The consensus was that other tasks such as writing a bestseller (2049) or carrying out surgeries (2053) wouldn't be quite so imminent. Interestingly, the experts (who answered in 2015) predicted that AI would not surpass humans at Go until 2027 — yet that's already happened. This suggests the sobering thought that in general their predictions may have been far too conservative against AI.
Still, even if we go with the estimates the experts provided — and these were attendees of two of the most significant AI events in 2015 — there is a 50% chance that AI will surpass human intelligence in all areas within about 45 years. AI researchers from Asia think it will happen in 30 years, while AI researchers in North America think it won't happen for 74 years.
Getting Ready For AI
Regardless of which estimate is accurate, there isn't much time left before AI will be capable of taking over any — if not every — job that exists right now. This means that the time to address the potential fallout from this change to our economy and culture is right now. Canada, as well as many other countries, is preparing for the automation age by investing in training and education. Many here in the U.S. have called for more investment into affordable education for the same reason.
Various nations and private companies are also planning or trialling universal basic income (UBI) programs, including Canada and a German startup in Berlin. Various tech leaders such as Mark Zuckerberg and Elon Musk support UBI in the U.S., and see it is as the only possible future in light of automation. While there is no consensus around the world about how to prepare for the rise of AI, there can be no reasonable doubt that now is the time to begin planning and taking action.