Meta AI searches made public - but do all its users realise?
Imagine if your internet search history was suddenly available for everyone to see. This is the reality for some users of Meta AI, who may not even be aware that their interactions with the artificial intelligence tool are being shared on a public feed. As a result, their search queries and the responses generated by Meta AI are visible to anyone who comes across them.
The Privacy Concerns
According to internet safety experts, this poses a significant user experience and security risk. Some posts made by users can easily be linked back to their social media accounts, potentially revealing sensitive or personal information. While Meta AI claims that chats are private by default and users have the option to make their posts public, the implications of sharing such information are alarming.
Before a post is shared publicly, a warning message does appear, advising users to avoid sharing personal or sensitive details. However, the extent to which users comprehend the implications of their actions remains uncertain. It's unclear whether users fully understand that their searches are being posted on a public "Discover" feed on the Meta AI app and website, where they can be linked to other accounts through usernames and profile pictures.
Examples of Revealed Searches
The BBC discovered several instances where users posted queries on Meta AI that raised privacy concerns. Some users uploaded images of exam questions, seeking answers from the AI. Others explored topics related to their gender identity and potential transitions.
Shockingly, there were also searches for images of scantily-clad women and anthropomorphic animal characters. One particularly concerning search linked back to an individual's Instagram account, requesting Meta AI to generate an image of a character in revealing attire. These instances highlight the inadvertent sharing of personal information on a platform that users may not realize is public.
Meta AI's Features and User Control
Meta AI, introduced earlier this year, can be accessed through social media platforms like Facebook, Instagram, and WhatsApp. It offers a standalone product with a public "Discover" feed, allowing users to explore how others are utilizing AI. Users can adjust their settings to keep their searches private.
Expert Insights
Rachel Tobac, CEO of a US cybersecurity firm, voiced concerns about the potential risks posed by Meta AI's public sharing feature. She highlighted the lack of alignment between user expectations and the platform's practices, underscoring the need for enhanced user awareness and control.
As users continue to engage with Meta AI, it is crucial for them to evaluate the implications of their interactions and exercise caution when sharing personal information on public forums.
