Elon Musk’s X is officially entering the generative AI race, with xAI’s first chatbot, named “Grok” now in limited testing, with a view to a wider launch in the near future among X Premium+ subscribers.

X announced its X Premum+ tier last week, which, for $16 a month, gives you access to all of the X Premium features, and also eliminates ads for your timeline.

And now, you’ll also get access to Grok, which, at this stage, looks a little like this:

So what makes Grok special, and differentiates it from, say, ChatGPT, and other AI chat experiences?

The main variation of Elon’s AI system is that it’s, in Musk’s own words, intended to be “anti-woke”, with less censorship of potentially sensitive topics. 

Elon has voiced strong criticism of ChatGPT’s perceived censorship of certain answers on certain topics, so he launched his own AI project to be a better representation of “truth”, in his own definition and experience.

So how will it do that? 

Elon believes that Grok will be a more accurate source of truth, particularly on real-time news topics, because it’s sourcing its data from X, with X’s updated user agreement allowing it to use people’s posts to fuel the system.

X has also been working to cut off other generative AI projects from its API, especially OpenAI, who Musk has personal grievances with. As a result, Grok, and X’s other AI projects will be among the limited few that have full access to the full X firehouse of information.

Which, at least in theory, should be a significant advantage in some respects. But at the same time, X has also been losing traction as a key news discussion platform, due to Musk’s various changes at the app, which have included a range of updates designed to undermine “mainstream” news outlets, which, in turn, has seen many journalists migrating to other platforms.

And any change in this respect is significant, because while X has 244 million daily active users, very few of them actually ever post anything in the app.

According to X, 80% of its users are in “read only” mode, and never post, nor engage in any way. That means that X’s entire input of 100 million unique posts per day is down to 49 million users.

So X’s active posters are only posting, on average, two posts per day, though that doesn’t account for replies (which are another 100m per day).

But it’s actually even less than that.

X recently reported that 42% of its users are sports fans, which suggests that a significant portion of those 200m posts and replies each day are also sport-related (otherwise X would have no way of qualifying this figure).

Assuming that not all sports fans post exclusively about sports, but many do, let’s say that 30% of X’s daily posts are about sports. That means that only that there are only around 70 million original posts per day in the app about non-sports subjects.

Which is a much smaller input stream than it would initially seem, and hardly indicative of global trends or opinion. And with the most active X users being X Premium subscribers, many of whom are now making money from their posts, the skew of those 70 million posts is also likely to be heavily aligned with Elon’s own opinions on various divisive topics, including COVID, the war in Israel, media misinformation, U.S. politics, and more.

In other words, Grok is likely to be a right-leaning misinformation machine, basing its responses on Elon’s army of “citizen journalists”, who now share whatever they want in the app.

But wait, aren’t they being fact-checked by Community Notes?

Well, not really. Community Notes has proven beneficial for some disputed claims in posts, but the system is already being manipulated, while each note also requires consensus from contributors with opposing political leanings, which means that on many of the more divisive topics, relevant Notes are never displayed, because such agreement will never be reached.

So certain misinformation is effectively immune from Community Notes. Which is likely how Elon wants it, but it doesn’t exactly bode well for a question and answer-based service that’s focused on sharing up-to-date news updates. 

But Elon says that Grok will “have a little humor in its responses”, and it’ll be sarcastic, because it’s trained on X posts. So it should be fun, even if it is also potentially offensive, harmful, defamatory, dangerous, etc.

But Elon is convinced that he has the formula for helping to spread truth, away from media and government lies, and that this, eventually, will fuel movements of citizens that share the real happenings of the world.

What could possibly go wrong?