AI won't take as many jobs as expected

San Francisco : While researchers around the world expect job losses due to artificial intelligence (AI) soon, a new study has said that AI might not take as many jobs as expected.

A recent study conducted by MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) investigated whether AI could perform tasks more efficiently than humans and whether it was cost-effective for businesses to replace human labour with AI.

The study took into consideration the broader implications of AI implementation in the labour market. The researchers found computer vision AI can currently automate tasks that makeup 1.6 per cent of worker wages in the US economy, excluding agriculture.

However, only 23 per cent of those wages, equivalent to 0.4 per cent of the entire economy, would be cheaper for companies to automate instead of hiring human workers at current costs.

"We find that at today’s costs US businesses would choose not to automate most vision tasks that have “AI Exposure,” and that only 23 per cent of worker wages being paid for vision tasks would be attractive to automate," the authors said.

"Overall, our findings suggest that AI job displacement will be substantial, but also gradual -- and therefore there is room for (government) policy and retraining to mitigate unemployment impacts," they added.

The study included examples of "vision tasks" that AI could achieve including analysing images from hospital diagnostic equipment or examining trays to ensure they contain the correct items. Researchers surveyed employees to determine the share of their tasks that could be accomplished by computer vision. They subsequently developed models to assess cost-effectiveness. They discovered that using AI visual detection to replace workers would rarely be worthwhile.

"We find that the median employee works in a firm where close to none of the vision tasks are cost-effective to automate. Even a firm with 5,000 employees, i.e., larger than 99.9 per cent of firms in the US, could only cost-effectively automate less than one-tenth of their existing vision labour at the current cost structure," the researchers stated.

The study acknowledged that the cost of AI will decrease over time, but the authors don't think it will do so as quickly as some have suggested. They also predicted that it would take some time for AI to have a major impact on these kinds of professions.