Why Meta added AI image creation to Messenger
It's not just a party trick
Have you played with Meta’s AI image creation in Messenger, WhatsApp or Instagram? It may seem like just a fun party trick (and it is fun!), but it’s also a serious strategic play.
Ok, you got me, I haven’t seen this yet. What did they build?
Embedded into Messenger, WhatsApp and Instagram is now the ability to either use AI to modify your own photos or imagine images from scratch. It’s really powerful and really fun to play with. Here’s a screen capture that demos how I took a photo of my son standing in the living room and transformed it into a picture of him on a tropical beach.
So why did they do this? How is it strategic?
For Meta this new feature serves 4 strategic purposes:
Shows progress to investors
Acclimates the public to using AI
Increases platform engagement & retention
Generate training data
First, Meta has been investing billions into AI research and development. In their last earnings call, Meta raised their estimated annual capital expenditures guidance to $38 billion (see CNBC). Much of those expenses are associated with AI development. Yet, what do they have to show for it? They have built Meta Quest, Ray-ban Meta and Llama, their open source AI - which is a lot, but it’s not readily accessible. Adding AI directly into their prolific consumer apps provides tangible evidence to the public and investors that Meta is an AI leader.
This move also begins to acclimate the public to using and accepting AI as part of their normal, daily interactions. This will be crucial for Meta since they are so heavily invested in AI. For those investments to pay off AI can’t be an abstract thing that only techies interact with or understand, it needs to be something commonly used and accepted. Meta is seeking to change people’s behavior and attitude - something exceedingly difficult to accomplish as part of a product strategy. Embedding AI into apps that people use daily (hourly!) in a non-threatening accessible (and fun!) way seems to be one part of that strategy.
There’s also a direct monetization pay-off to acclimating the public towards using AI. Meta has just made it cheaper for advertisers to generate engaging images for paid social media posts. Embedding AI image creation within Meta’s suite of messaging apps means that the primary use case will be people creating fun images and sharing them. Each image contains a little “Imagined with AI” badge - which over time people will grow accustomed to seeing and accepting. This acceptance will then lead advertisers to use Meta’s AI image creation for paid posts. AI will make it faster and cheaper for advertisers to create images for those posts.
The third strategic reason behind this move is driving platform engagement and retention. Daily Active People (DAP) is the first metric the company reports outside of financials (see their Q3 earnings). Intuitively this makes sense, DAP is traffic, traffic means ad impressions, advertising drove 98% of Meta’s Q3 revenue ($39.88B out of $40.58B). Without innovations like these in their core apps, people will migrate to other apps that deliver more entertaining experiences.
Lastly, adding these capabilities is a huge win for Meta’s AI product and research teams. There is an enormous potential for learning. The teams will study the types of prompts people type (and how accurately the AI responds) and user behavior with AI itself. This in turn will improve the quality of the AI product from both an algorithmic, technical perspective as well as from a user experience perspective. By understanding how their customers will act around an AI agent, Meta’s PM and Design teams can craft more intuitive, comfortable experiences.
In conclusion: it’s a smart strategic play masquarding as a fun party trick. But that’s been Zuck’s style dating back to 2004.