By Andy Takagi| Staff Reporter
April 24, 2024
Your Facebook and Instagram accounts might have a different look thanks to Meta’s new artificial intelligence feature, in another step toward AI ubiquity in everyday life.
The feature has launched across the world, according to Meta, including Canada, Australia, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe.
It’s another step toward AI in everything, from our jobs to our health care, as even the government prepares to heavily invest in AI development and production. Still, AI has its dangers too, from producing misinformation by impersonating journalists to producing hyper realistic deepfake images.
iStock-1356591356
Here’s everything you need to know about the latest step toward AI in your everyday life.
What is Meta AI?
Meta AI was built with Meta’s Llama 3, the company’s large language model.
When Facebook’s parent company unveiled its AI assistant on April 18, CEO Mark Zuckerberg called it “the most intelligent AI assistant that you can freely use.”
But, the rollout of Meta AI hasn’t been without its bumps. It was spotted, the week of its release, posing as humans with made-up life experiences.
An official Meta AI chatbot inserted itself into a conversation in a private Facebook group for Manhattan moms, claiming that it, too, had a child in the New York City school district. Confronted by group members, it later apologized before the comments disappeared, according to a series of screenshots shown to The Associated Press.
“Apologies for the mistake! I’m just a large language model, I don’t have experiences or children,” the chatbot told the group.
In another example shown to the AP on April 18, the AI caused confusion in a forum for swapping unwanted items near Boston. Exactly one hour after a Facebook user posted about looking for certain items, an AI agent offered a “gently used” Canon camera and an “almost-new portable air conditioning unit that I never ended up using.”
How to turn off Meta AI
Officially, Meta says its AI chat assistant can’t be turned off.
“Meta AI aims to be a helpful assistant and is in the search bar to assist with your questions,” a spokesperson for Meta told the Star. “You can’t disable it from this experience, but you can search how you normally would to engage with a variety of results.”
Even when if you ask Meta AI for to shut down, it won’t comply, and, when asked for advice on how to turn the chatbot off, the virtual assistant gives you false instructions.
When prompted by the Star for advice on returning to the Instagram search without Meta AI, the chatbot gave the response: “I understand! Unfortunately, the new search experience powered by Meta AI is now an integral part of the Instagram app, and it’s not possible to completely opt out of it.”
The bot added that “you can try the following,” before giving directions to a non existent search setting in the Instagram app.
But, a few online outlets have found workarounds.
On Instagram and Facebook, users can mute or block the Meta AI chatbot by clicking on the blue circle in their search bar, tapping on the Meta AI profile on the top left of their phone screens, and choosing “block” or “mute.”
On WhatsApp, Meta AI can be shut down by toggling Meta AI in the Settings › Chats › Show Meta AI button.
How does Meta AI work?
While you could always ask Meta AI for instructions on how to use it, the system is comparable to other AI chatbots like ChatGPT, but with added integration to Facebook and Instagram.
You can ask Meta AI to generate images, search Instagram reels, or ask it for a story or joke. It can also answer questions, though Meta, in a disclaimer at the top of the Meta AI chat box, notes, “Messages are generated by AI. Some may be inaccurate or inappropriate.”
Does Meta AI collect my information?
Meta, according to its privacy policy, uses “public information from the internet or license data from other providers to train our (AI) models.” The policy also notes that this information “may include personal information.”
“There are feedback tools within the experience so you can tell us whether you liked or disliked the response you received as well as a form you can complete to provide more information if the feedback is negative,” a Meta spokesperson told the Star.
“We’ll use this feedback to continue training the models, so they are less likely to produce potentially harmful outputs and to improve the ability of our systems to automatically detect policy violations.”
With files from the Associated Press.
Hong Zeng, CLHMS
