WhatsApp Pushes ‘Optional’ AI Feature Users Can’t Disable

WhatsApp claims that its recently added AI capability is “entirely optional” even though users cannot disable it from the app.

The ubiquitous blue circle with pink and green splatters in the lower right corner of your Chats screen is the Meta AI logo.

It has attracted notice and annoyance from users who are unable to remove it from the app, but interacting with it launches a chatbot that is intended to respond to your inquiries.

It comes after Microsoft’s Recall function, which was an always-on tool until the company opted to let users turn it off after receiving criticism.

Whatsapp thinks giving people these options is a good thing and they are always listening to feedback from their users,” WhatsApp stated to the BBC.

The ‘channels’ and’status’ features are comparable to other permanent options in the app, according to the business.

It occurs in the same week that Meta revealed an update to Instagram’s teen accounts feature.

The company disclosed that it was testing AI technology in the US that is intended to identify teen accounts on the site who have made false claims about their age.

The ‘channels’ and ‘status’ features are comparable to other permanent options in the app, according to the business.

It occurs in the same week that Meta revealed an update to Instagram’s teen accounts feature.

The company disclosed that it was testing AI technology in the US that is intended to identify teen accounts on the site who have made false claims about their age.

You might not be able to utilise it just yet if you are unable to see it.

According to Meta, the function “may not be accessible to you yet, even if other users in your nation have access” and is currently only being made available in a select few countries.

‘Ask Meta AI or Search’ is an option that users can use in addition to the blue circle on the top search bar.

Instagram and Facebook Messenger, both of which are owned by Meta, also have this feature.

Llama 4, one among the massive language models used by Meta, powers its AI chatbot.

There is a lengthy message from Meta outlining what Meta AI is and that it is “optional” before you ask it any questions.

“Can answer your questions, teach you something, or help come up with new ideas” is what WhatsApp claims Meta AI “can do.”

When I asked the AI what the weather was like in Glasgow to test the feature, it gave me a complete reply in a matter of seconds that included the temperature, humidity, wind, and possibility of rain.

Additionally, it provided me with two links for additional information, but this is where it encountered issues.

While one of the links was pertinent, the other attempted to provide me with further weather information for Charing Cross, which is the London train station rather than the Glasgow site.

According to users on X, Bluesky, and Reddit, people in Europe are not happy with it as of yet. Polly Hudson, a columnist for the Guardian, was one of those who expressed their annoyance at not being able to turn it off.

Dr. Kris Shrishak, a privacy and AI expert, was also quite critical, accusing Meta of “using people as test subjects for AI” and “exploiting its existing market”.

He stated to the BBC that “no one should be forced to use AI,”

“Meta has utilized people’s personal information and pirated books to train its AI models, which is a deliberate breach of privacy through web scraping. 

“Now that the legitimacy of its strategy has been confronted in courts, Meta continues to search for other means to collect data gathered from individuals, and this feature might represent one such source.”

According to an Atlantic investigation, Meta may have used LibGen, or Library Genesis, to access millions of illegal books and research articles in order to train its Llama AI.

Author organizations in the UK and other countries are launching campaigns to get governments to step in, while Meta is presently defending a lawsuit filed by several authors against the exploitation of their writing.

Regarding The Atlantic report, a Meta representative declined to comment.

The messenger bot “can only read conversations people share with it” when you initially use Meta AI in WhatsApp.

“Meta can’t read any other conversations in your personal chats, as your personal messages remain end-to-end encrypted,” it reads.

The BBC was informed by the Information Commissioner’s Office that it will “continue to monitor the implementation of Meta AI’s technology along with the utilization of personal data within WhatsApp” in the meanwhile.

“Personal information drives much of AI innovation because individuals need assurance that organisations are using their information wisely,” it stated. 

Dr. Shrishak advises users to exercise caution. He also stated “When you send messages to your friend, end to end encryption will not be affected.”

“Every time you use this feature and communicate with Meta AI, you need to remember that one of the ends is Meta, not your friend.”

The IT giant also emphasizes that you should only distribute content that you are aware others could find useful.

“Don’t share information, including sensitive topics, about others or yourself that you don’t want the AI to retain and use,” it states.