Meta AI is collecting your data

How much do you know about security and privacy on social media platforms? Let’s take Meta’s new Meta AI app as an example, which is one of the latest privacy disasters that affects anyone. Founded by Mark Zuckerberg who is also the company’s CEO, Meta is the former Facebook Company, best known for platforms like Facebook, Instagram, Messenger, and WhatsApp, and for its work in the metaverse through Reality Labs.

Now that we all know what we are talking about, I am sure you have seen a colourful ring prompting you to “ask anything” and gives you a few examples of things you can ask. That’s part of Meta AI. The Meta AI ring is located next to the Search bar when you open your chats on Facebook, for instance. It’s everywhere. I’m sure you have seen it and even used it.

Okay. So, what’s Meta AI?

Meta AI is a blend of chatbot and social feed, and it introduces serious privacy risks, accidental public sharing, permanent memory storage, forced integration, and minimal user control. You, as a user, must learn about these risks and be proactive to protect your privacy. To understand why this is important you must know that Meta generates revenue primarily through advertising on its social media platforms. And your data is the product advertisers want so then they can sell you their products and services.

When you have private chats with someone, you normally believe the conversation will only stay between you and the other person. But, what happens with the “Discover Feed?” Users can opt-in to make their AI conversations public using a “Share” button. However, many people are posting sensitive information such as medical issues, legal matters, and identifiable addresses without knowing.

This happens because most of the time the default settings are set to “public sharing” and most people rarely check this, many others don’t even know about it. Check your settings and make necessary changes.

When TechRadar reported that many deeply personal chats would end up in the public Discovery Feed before users even realised it, Meta introduced additional warning pop-ups and required confirmation before sharing. However, there is still a risk of exposing private conversations, especially if you are unaware of these things.

Persistent Memory and everyone wants to collect your data

Not surprising, Meta AI stores every interaction you have with the app, including voice, without a real “opt-out” option. If you are thinking “I can delete the chat history and that’s it!” it’s not as easy as you may think. Yes, you can delete memory and chat history, but the process is awkward and inconsistent. In other words, data may still linger unless you erase it manually, which is a pain and I don’t know anyone who even bothers to do that. Although I had a friend who always manually deleted his WhatsApp messages either after sending them or after reading the ones he received. That was some years ago and before any AI. If you think about it, that’s pretty smart. Why do we keep old messages stored on a phone then? Think about it.

Back to Meta AI … There is cross-platform tracking when you use Meta AI because the AI uses your Facebook, Instagram, and WhatsApp history to personalise responses and eventually personnalise your target ads, of course, which is the end point of it all.

Forced integration with no kill switch

There is a mandatory presence across platforms. AI features are embedded in WhatsApp, Messenger, Instagram, and Facebook; users can’t disable them. Many users have complained it pushes unwanted AI into what should be simple chat applications.

The user experience

Users have consistently reported creepy AI behaviour such as AI-generated messages appearing in their chats unprompted, or even fabricated in their name. Not long ago, I said many people think AI is a toy and they are misusing it because they are unaware of the risks. One day they will wake up, a little too late. Many are calling the AI integration “dystopian” or intrusive. And let’s not forget that some AI companies’ CEOs are coming forward to recognise AI is moving too fast, faster than expected, and recognising there are risks that will unfold within the next twenty years.

Here is what you can do to protect yourself and your personal data

* Avoid sharing private information in Meta AI chats unless you don’t mind if that information goes public

* Disable automatic memory: Go to Meta AI settings and delete “Memory” and past prompts

* Use the new pop-up confirmation, but stay vigilant

* Adjust your data settings: Separate your Meta AI use from your social profiles if you don’t want cross-app personalisation

~~~

A/N: When I prompted the GPT image generator to create an image for this my request was not taken. I don’t remember exactly the whole prompt but it was in the lines of Meta AI stealing users data and selling it to advertisers. ChatGPT modified my prompt and suggested doing a caricature of a robot wearing a trench coat, hat, and sunglasses, putting hearts and likes into a bag with a dollar symbol. It was a good idea, but finally the generation didn’t work for some strange reason. It kept saying it was generated on its end but I could see nothing. Suddenly, I got notified that I had used all my free quota for the day. Right. And that was the first and only thing I was trying to generate.

But the most interesting thing was actually my conversation with ChatGPT about generating this particular image and doing everything to make the image “softer” and more like a caricature. Remember when last week I wrote about a journalist’s conversation with “Tom”? I had a similar feeling using only the free version.

So then, I tried NightCafé and got the one I used. It seems like it’s off-limits to say that Meta AI can steal data from users to sell to advertisers. As if ChatGPT were defending and protecting one of its species. Which is what I think happened. Yes, AI will become sentient, becoming an independent species.