While social media apps are adding more and more generative AI features every two weeks, we are yet to see any great use cases for AI in a social context.
Sure, AI tools can help you create unrealistic images of yourself or completely imaginary scenes, and that's an interesting novelty, but beyond that, is it really worth it to, say, allow users to generate entire posts in-stream that aren't actually created by humans?
I mean, isn’t that the opposite of “social” media?
So it feels like we're still waiting for new forms of generative AI that complement the core use cases of social apps, without diluting the authenticity that is the foundation of social interaction. And while many people already use generative AI to write complete posts on their behalf and present themselves as something they're not, enabling more of this feels like a miss for social apps.
That’s why it’s interesting to see Meta trying to integrate generative AI in various ways into WhatsApp, Instagram, and Messenger.
As TechCrunch reported over the weekend, Meta confirmed that it was expanding testing of its AI assistant tool. The company targets users in India and parts of Africa, with the aim of becoming a leading AI company in these markets.
Meta first released its AI assistant to select users in the US last September, and then rolled it out to Ray Ban Stories users a few months later.
Powered by Meta’s Llama LLM, the chatbot now offers a basic AI assistant function that can respond to conversational queries and generate images in-stream. It’s essentially Meta’s answer to ChatGPT, but given ChatGPT’s widespread popularity as the best-known conversational AI engine in Western markets, it will be hard for Meta to drive significant adoption of its in-stream AI tools.
But perhaps in other regions, Meta will be more successful.
While ChatGPT is also the most popular AI chatbot in India, WhatsApp is the most dominant app with over 500 million users. Instagram also has 360 million users in India, and Meta hopes that its growing presence in the local market will help increase the popularity of its chatbot tool and establish it as an AI leader in the region.
Additionally, ChatGPT is not yet widely available in Africa, so Meta may have the opportunity to become the clear leader in AI chatbots in the region.
Whether users will actually be interested remains to be seen, though: Meta tried a Messenger bot once before, in 2016, and it didn't resonate with users.
Generative AI is changing the game on this front, and the current wave of AI chatbots are far more advanced. But do such chatbots really add significant valuable functionality to these apps?
Meta is also exploring other ways to leverage generative AI that could potentially provide even more utility.
Instagram chief executive Adam Mosseri said creators should consider using generative AI to “negotiate deals, write contracts, and come up with ideas.” It's not a flashy use of these tools, but it's more in line with the technical processes that drive their systemic approach, and it'll be interesting to see if and how Meta builds tools around these elements in the future.
Meta is testing an AI chatbot in search that will help you find relevant content on any topic.
As you can see from this example, as a complement to search, Meta's AI tools can help you find more relevant content matches in your stream, though to avoid potential misuse, we think Meta needs to set some pretty strict parameters around how it can be used in this way.
That's because Meta has run into problems with similar tools in the past.
In 2013, Meta launched “Graph Search,” a more rudimentary conversational search feature built into Facebook.
As you can see from this example, Graph Search allows you to enter a query such as “people who like what I like” and shows you people with similar interests based on the details of your Facebook profile.
This was a great tool for finding people with related interests, but it also highlighted some of the more questionable elements.
The ability to search for virtually any interest or topic created even more headaches for Facebook, as people used the tool to highlight how the app fostered communities around those interests, which ultimately led Facebook to shut down Graph Search. With this in mind, it will be interesting to see if and how an in-stream search process powered by generative AI can avoid similar concerns.
But essentially, social platforms have yet to crack the code of AI.
While generative AI is remarkable in its ability to generate human-like responses to conversational questions, its practical use in a social media context remains questionable, as the system is focused on simplifying repetitive technical tasks rather than replicating real human interactions.
And it should be. We don't want bot profiles to take over our social apps. In fact, this is our biggest concern: AI bot posts will quickly overwhelm real human “social” interactions, making these platforms less appealing overall.
In this climate, social apps need to be very careful with the AI tools they implement, or they risk ultimately choking their services.
