Meta is facing backlash over privacy concerns surrounding its AI-powered smart glasses. According to an investigation by a Swedish newspaper, contractors in Kenya reviewing AI data may have seen highly personal footage captured by the glasses. This includes bathroom visits, intimate moments, and credit card information. These allegations have prompted legal action and sparked debate around AI systems. Tech and cybersecurity analyst Ritesh Kotak explained that when setting up the device, users are asked if they want to improve the product experience by sharing information with third parties. He added this is why he encourages people to read the terms and conditions. “You don’t really know what you’re agreeing to,” he said in an interview with CTV News on Monday. “And what you’re agreeing to is that any of the recordings that are stored in the cloud can essentially be sent to a third party for advertising, for analysis, to improve the product and that could be anywhere in the world.” Though many people think AI systems are trained by artificial intelligence, Kotak said this is not actually the case. He explained it’s actually human beings training the algorithm, describing it as “armies of people” looking at the data to ensure it’s correct. “Every time that there is an issue or some sort of transcription or you report something that’s wrong with the result, there’s human beings that actually review it to see A. what went wrong and B. if it is getting it right,” he said. Watch above to find out more about how AI systems are trained, and whether Kotak thinks the news about the AI smart glasses will hurt sales. • With files from CTV’s Maralee Caruso
