Innovation

Experts say there's a 50% chance AI will outperform humans in every job in 45 years

Yale and Oxford University researchers recently surveyed a host of AI experts regarding their thoughts on the timeline for AI development, and what jobs are at risk of being automated.

There's a 50% chance that artificial intelligence (AI) systems will be able to perform every single task better than their human counterparts in the next 45 years, according to researchers from Oxford and Yale University. The research also gave the same odds that AI automate all human jobs in the next 122 years as well.

The research, titled When Will AI Exceed Human Performance? Evidence from AI Experts, was submitted and later revised in late May 2017. Its purpose is to help anticipate the innovation in AI and machine learning for the purpose of developing public policy.

In terms of how soon AI will begin outperforming humans in specific tasks, the research laid out a short timeline. AI will best humans at language translation by 2024, writing high-school level essays by 2026, truck driving by 2027, retail work by 2031, writing a bestselling book by 2049, and working as a surgeon by 2053, the research said.

SEE: Inside Amazon's clickworker platform: How half a million people are being paid pennies to train AI

The baseline the researchers used for their survey was high-level machine intelligence (HLMI) which the report said "is achieved when unaided machines can accomplish every task better and more cheaply than human workers." While the aforementioned 45 years to HLMI was given a 50% chance, the research also noted a 10% chance that it would occur in nine years.

The researchers collected data from some 352 researchers who responded to their survey. The 45 year mark was an aggregate number, but predictions varied widely by geographic region, the research said. For example, surveyed researchers from Asia predicted HLMI in 30 years, while North American researchers pegged it at 74 years.

aiworker.jpg
Image: iStockphoto/Jirsak

Additionally, 67% of those surveyed said AI had progressed faster in the second half of their career, pointing to a recent growth in the field. However, while positive outcomes in HLMI advances are possible, there are also potential "catastrophic risks" and work should be prioritized to minimize the risks of AI, survey respondents said.

While the findings may seem extraordinary, it's worth noting how limited current AI and machine learning systems are. There is still substantial human involvement in wrangling, cleansing, and processing the data needed to power these algorithms, and most AI is extremely narrow in its focus. For example, Amazon had to hire an army of workers to feed data into its systems and train its AI, because the system itself couldn't keep up.

A previous Oxford study claimed that 47% of jobs were at risk of becoming automated in the next 20 years, which helped to highlight the concern of many modern workers. Still, other experts have claimed that AI and automation will primarily impact specific industries, while potentially creating more jobs in the process.

The 3 big takeaways for TechRepublic readers

  1. Yale and Oxford University researchers surveyed AI experts who said there's a 50% chance AI will outperform humans in every job in 45 years.
  2. The research also said that, in the next 122 years, there is a 50% that AI will automate all of those jobs as well.
  3. There are promises to AI advances, but also potential for "catastrophic" risks, survey respondents said, and steps should be taken to address these risks.

Also see

About Conner Forrest

Conner Forrest is a Senior Editor for TechRepublic. He covers enterprise technology and is interested in the convergence of tech and culture.

Editor's Picks

Free Newsletters, In your Inbox