Privacy nightmares. Pervert glasses. These are some of the not-so-flattering ways detractors describe Meta’s Ray-Ban AI glasses.

Now, we’re nominating another hellish descriptor into that list: dysmorphia accelerator.

On Friday, Meta announced a host of new features for the smart spectacles that include the ability to track and give advice on everything a wearer eats — which sounds poised to lead a bunch of people into horrendous eating disorders.

In an upcoming update, Meta says users will soon be able to use a voice command or a quick photo to log the food in front of them, and the Meta AI will automatically “extract key nutritional details” and add them to the wearer’s Meta AI app.

You can ask the AI questions like “What should I eat to increase my energy?” the announcement states, and it will give an answer informed by the user’s food log and personal goals.

“Over time, your food log powers increasingly personalized insights that get more useful, helping you make healthier, more informed choices,” the announcement reads.

The most eyebrow-raising detail comes when Meta says that, down the line, it’ll deploy an even more ambitious functionality: the AI automatically logging your food without you even having to ask. This will be possible because the Ray-Bans will “understand what you’re eating,” giving wearers “even richer, more personalized nutrition insights without having to remember to log every meal.”

Putting aside all questions of privacy — presumably the glasses will have to be constantly recording to automatically log your food, even in public spaces like restaurants — this just sounds like a bad idea. The reason calorie and nutrition tracking is a chore is because it takes a combination of research, guesswork, and measurements. How are AI glasses supposed to accomplish all that, providing even remotely accurate predictions when it both has to infer details like what the food it’s looking at is, what it contains, and portion sizes?

It’s also plain to see how the AI could give bad advice or be used to reinforce someone’s existing neuroses around food. The heaps of reports of so-called AI psychosis are testament to that. (We submit this story we reported on a father who, following the Meta AI glasses advice, ended up searching the desert for aliens.)

What if the question is along the lines of, “will skipping this meal make me skinny?” or “am I too heavy?” — or what about the opposite: someone trying to bulk force-feeding themselves into a heart attack off of an AI’s advice? We’ve already seen instances of chatbots like ChatGPT giving teens advice on how to go on starvation diets, so none of those possibilities seem far-fetched.

The feature is set to debut this summer, and will only be available to 18 years and older in the US. That might indicate that Meta, seemingly, is aware of the risks.

More on AI: Meta Workers Say They’re Seeing Disturbing Things Through Users’ Smart Glasses