WhatsApp tests AI content labels, unofficial app warnings
WhatsApp week's beta roundup covers Liquid Glass menus, mandatory AI labels, and a new warning against fake WhatsApp Web clients
WhatsApp's latest round of beta features leans heavily toward safety and transparency, alongside a design update that finishes a job it started months ago. Here's everything spotted in testing this week.
The reaction tray and context menu on iOS are getting the same translucent Liquid Glass treatment WhatsApp already applied to the chat bar.
Reacting to messages works exactly as before; only the visual layer changes, with transparency effects replacing the previous solid background and the interface adapting to whatever appearance setting a user has chosen.
WhatsApp says drawing tools and the video player are next in line for the same treatment.
AI content labels can now be added by channel admins on both the Android and iOS versions of the app in all updates, including media generated and edited with the help of AI technology.
The label cannot be deleted once it is added because it acts as a disclosure, not as a toggle, and appears in the same corner of the message box where view counts and timestamps are present.
It does not matter how the content was created, whether through ChatGPT or Gemini or through Meta AI; all of them have to follow the disclosure policy. In many countries, the labelling has become mandatory.
A new warning is being tested by WhatsApp, which indicates the unofficial WhatsApp Web application on the Linked Devices screen.
The reason why this is a problem is that these third-party applications work as an intermediary between the user's account and WhatsApp's server, which means that the messages and media may travel through applications that have not been developed by WhatsApp.
The warning has been issued after it was discovered that 200 people in Italy had become victims of government-sponsored spyware due to the distribution of the fake application through phishing links.
Rounding out the week, WhatsApp is testing notifications that alert users to unfinished draft messages a minute after they leave a chat; tapping the alert reopens the conversation directly.
The company also continued rolling out its third-party agent chat feature and the expanded four-message pin limit, both still in staged release to beta testers.
-
Berlin reportedly launches crisis response after hackers publish stolen data
-
OpenAI’s GPT-6 Astra shows off news skills in music, games and complex tasks after launch
-
Did you know Microsoft spent $1bn on phones it killed in 48 days?
-
OpenAI admits it needs new rules for rogue AI incidents
-
Sam Altman’s Apple fandom gets complicated as company sues OpenAI
-
Hackers leak nearly 6TB of sensitive Berlin government data
-
Elon Musk picked a fight with Chess.com, it didn't go well
-
OpenAI plans new AI misalignment reporting framework after German wiki incident