A bit too smart.
Meta
The Ray-Ban Meta smart glasses are a step toward the tech-filled future we all imagine, but perhaps we'll need to get used to giving up some privacy in the process.
Meta has confirmed that it can use the images and videos you share with Meta AI to train the system.
"[I]n locations where multimodal AI is available (currently US and Canada), images and videos shared with Meta AI may be used to improve it per our Privacy Policy," Meta policy communications manager Emil Vazquez told TechCrunch.
There is a silver lining: whatever you capture with the glasses and don't analyze with Meta AI is safe and won't be used to improve AI, but once you let it work on the images, they fall under a different set of policies. In short, if you don't want Meta to train AI on your media, don't use its multimodal AI features.
However, Meta's policies also say the glasses store your audio recordings to "train Meta's products":
"Based on your settings, we may also use stored recordings and transcripts of your voice interactions and related data about your voice interactions to help improve Meta's products. When storage of voice interactions for product improvement is enabled, we use machine learning and trained reviewers to process information to improve, troubleshoot and train Meta's products."
As TechCrunch noted, Meta trains its Llama AI models on what people post on Instagram and Facebook, but this adds another, less noticeable, dimension to the strategy.
Smart glasses are a great invention, but you should be careful with how you use them. Read the original article here and join our 80 Level Talent platform and our Telegram channel, follow us on Instagram, Twitter, LinkedIn, TikTok, and Reddit, where we share breakdowns, the latest news, awesome artworks, and more.