Many within the online technology community are up in arms about the latest move by Meta regarding Facebook.
The company decided the future of social media is AI generated users. Here is a recent headline from The Financial Times:
So what is going on?
This move is leading many to espouse that Zuckerberg is trying to bring the Dead Internet Theory into being.
Actually, this shift is exactly what we have discussed for more than a year. In this article, I will dive into what Zuckerberg is up to and why I think this is the focus.
As always, there is a lesson for Web 3.0 contained in here.
Facebook: AI Generated Users Future Of Social Media
The Dead Internet Theory is the idea that the Internet will become a mechanism of automated activity and that humans will have a small part of it.
Actually, I do not believe it is a theory anymore. We know that more than half the Internet traffic is generated by computers. It is also something that is going to increase over time.
That said, it does not mean that human engagement declines. The total amount of online activity is skyrocketing. Certainly, it is driven mostly by computers. However, human engagement is also increasing as we do more online.
Thus, we have a situation where the growth in the amount of computer generated traffic is outpacing that of humans. The totals are simply shifting more in favor of the machines.
It also does not make the Internet dead.
Meta is applying this concept to social media.
So what is the company doing?
Meta is betting that characters generated by artificial intelligence will fill its social media platforms in the next few years as it looks to the fast-developing technology to drive engagement with its 3bn users.
The Silicon Valley group is rolling out a range of AI products, including one that helps users create AI characters on Instagram and Facebook, as it battles with rival tech groups to attract and retain a younger audience.
“We expect these AIs to actually, over time, exist on our platforms, kind of in the same way that accounts do,” said Connor Hayes, vice-president of product for generative AI at Meta.
This is what sent many off the deep end. Meta envisions a time whem AI characters will dominate their platforms. Users (human) will engage with these characters along with them interacting among themselves.
As crazy as this sounds, it is perfectly logical.
So what is Meta up to? Actually, they are doing exactly what we discussed in a number of articles.
Massive Data Generation
It always comes back to data. Meta found a way to generate a massive amount of it.
Consider what the data creation is with millions (billions) of AI characters all over Facebook and Instagram. Meta already is a treasure chest of data. However, with the hunger that training models have, this is not enough.
This is where the AI characters enter.
The massive explosion will be in synthetic data. We covered this topic and how there is signficant debate as to the value, long term, in training of models. If this is the case, Meta is also solving the problem in my view.
Regardless of which side of the debate about synthetic data is correct, there is one answer. The key is simply to get human feedback on the synthetic data. In other words, have the humans add to the synthetic data, helping to expand the meaning.
Obviously, this is what Meta seeks to do.
By developing characters, the goal is to get humans engaging with the AI generated content. As humans have "conversations", the feedback loop is constructed. AI and humans interact, providing Meta with enhanced synthetic data with regards to the value.
For that company, they are looking at doing it on a massive scale.
Developing Expertise
Quantum mechanics talks about the "observer effect". This is an important component as we move further into the AI world.
The problem with the major portion of the Dead Internet Theory is we come back to the tree falling in the woods. If no humans observe the interaction, what really took place?
For example, a conversation between two AI characters, if that is never read, what value does it have? At some point, the data generated will already be in the system. There is only so much that it can develop on its own. Naturally, this is much different from an AI agent autonomously handling tasks that go unoberserved. In that instance, we can see the results.
Meta is banking in human engagement. While many are quick to conclude the idea absurd, it actually is not. We are simply looking at the gamification of human interaction.
For example, if you want to learn about Einstein's physics, one could undertake the task by going to many different websites, papers, or taking a course. What happens when one can simply converse with Einstein himself (in digital form) to get the answers?
This concept could apply to any topic.
It Always Comes Back To The Data
When I look at moves made by these companies, I simply view it through the lens of data. What is being created or captured?
For the next half decade, that is all that matters. Even devices or services that offer other utility is nothing more than another data generation tool. Companies are going to look to keep developing these wherever possible.
For a company like Meta, there is no such thing as too much data. Some companies might get overwhelmed due to the costs. Meta has enough money to acquire whatever level of processing required to train all the data. In fact, like all of Big Tech, it is looking at a shortcoming in the next few years.
Basically, data has to go on another exponential growth curve. To achieve this, computer generated data is crucial. As Meta is showing, getting humans to engage with it is equally as important.
There will be AI characters doing everything under the sun. How about having Einstein, instead of physics, delivering the weather to you each day? That is something that will be possible.
Of course, to achieve that end, more data is required.
This is what I think Zuckerberg is focusing upon. He realizes that social media is nothing more than a pathway to data.