Meta is taking another step to make its feed recommendations and artificial intelligence responses more personally relevant, by incorporating data gathered from ad partners into its algorithmic mix.
Meta gathers data from ad partners to help personalize ad content, including information delivered via its Meta Pixel website code, which shares website visitor activity info back to Meta. That enables Meta to deliver more relevant ads based on external actions. Now, that same info will also ensure improved algorithmic recommendations and AI responses.
As explained by Meta: “We’re updating how we use information that other businesses already share with Meta. We already use this data — like games you play or purchases you make on other websites — to make the ads you see more relevant. In the future, we’ll use this information to personalize other parts of your experience, including the content you see in your Feed and AI responses.”
Of course, Meta is framing this as a benign update. The company used the example of a user who recently purchased a tent online, and who might now see more Reels about camping.
In some ways, this update is relatively limited in scope, and tame in data usage terms. But the more valuable element for Meta is likely the customization of Meta AI responses, which will now be able to incorporate more information about off-platform actions in order to deliver more relevant, targeted replies.
With this, Meta AI will know what websites each user has visited, what users looked at while they were there, what they bought and other details. That expanded data set will facilitate increasingly personalized AI responses, which is another step towards Meta’s vision of delivering “personal superintelligence” to everyone.
Imagine, now, a Meta AI chatbot that knows brand preferences and off-platform activity. That will help to shape increasingly customized AI bots that will feel more familiar to users. That knowledge and understanding can then be communicated through a conversational tone. In turn, many users could become more enamored with Meta’s AI bots, which will increasingly seem like they’re actually thinking, like they actually know users and like they can act as trusted friends.
But they’re not. The AI chatbots are not thinking, they’re not conscious and they’re only impressive because of the speed at which they can process massive amounts of data. That, in turn, can convince people that these AI tools are intelligent, and are delivering measured, considered advice.
That misinterpretation could lead to concerning impacts, as people become more reliant on AI companions, and increasingly convinced that these tools actually know them. However, AI systems are no more alive than Google search, which uses binary input to deliver estimated, calculated responses.
More data inputs will make chatbots seem more responsive, which could also see them veer into creepy territory. But the bottom line is that these tools only provide the perception of human understanding and consideration, a simulation of life.
Meta has sought to reassure users that these extended data inputs won’t impinge on existing permissions and/or settings.
Most importantly, Meta has specifically stated in its announcement that it is not collecting any new data as part of this update. Users will remain in control of how the company uses data to personalize their experiences.
To assist on this front, Meta is also merging its control tools to make it easier for users to manage how their personal information is utilized.
“We currently offer two settings that let people manage activity shared with us from other businesses: ‘Your activity off Meta technologies’ and ‘Activity from other businesses.’” Meta said. “Instead of maintaining two settings that cover similar ground, we’re streamlining our controls and will no longer offer the ‘Your activity off Meta technologies’ setting that lets you disconnect activity that businesses share with us from your account.”
It could be handy in some contexts, with Meta now able to provide more targeted recommendations based on a wider pool of data on users.
But it could also be intrusive, and the use of these expanded data sets within AI tools could also lead to more users becoming more reliant on AI bots as trusted advisors and companions.
Is that a good thing? For Meta it is. Unclear for everybody else.
Meta said that its control and data-use changes will first go into effect in the U.S., with other regions to follow in the coming months.
