As Elon Musk continues to re-angle his companies towards AI and robotics, you can expect X’s Grok chatbot to get even more focus moving forward.
Already getting prominent placement, right in the middle of the bottom function bar in the app, X is now developing a new way to use Grok, with a “Grok Analysis” button to fact-check any post in the app.
As you can see in this example, X is working on a new display of X posts when expanded, which will include large “Grok Analysis” and “Related Posts” buttons at the bottom of the UI.
Musk says that the analysis option will eventually provide AI fact-checks of any post, though it’s not quite ready as yet.
That could be because Grok itself seems to have trouble distinguishing fact from reality as yet, based on several Grok-originated trend highlights that have misinterpreted X posts about events.
Like Grok’s report that NBA star Klay Thompson went on a vandalism spree after a bad shooting night (throwing bricks), or that the sun had been “behaving unusually” sparking confusion (solar eclipse).
Or worse, that Iran launched a missile attack on Tel Aviv.
All of these false headlines were based on the Grok system misinterpreting sarcasm, or other messaging, from real time X posts, and it could be difficult for X to weed out false reports like this, based on how the system operates.
Because sarcasm will be a constant risk within X posts, while bot armies could also skew reports by mass posting about certain topics, in order to get such info to trend in the app.
In fact, X is even looking to ramp up the sarcasm with Grok’s coming “fun mode”, which is designed to provide humorous takes.
Hilarious. Grok’s done it again.
And while this is a design choice, I can’t imagine that having these types of reports amplified throughout the app is going to help Grok’s models in trying to ascertain actual truth.
Then again, that probably isn’t Musk’s aim, as he mostly seems to want X to reflect his own worldview, which includes a lot of random conspiracies and misinformation.
And nerdy jokes. So maybe, if you’re happy getting dime-store versions of comedic intellectualism, the “Grok Analysis” button will be for you.
If not, hopefully X is able to improve the Grok system to ensure accuracy in these fact checks. Because if it does actually end up working, it could be a very handy complement.
>>> Read full article>>>
Copyright for syndicated content belongs to the linked Source : SocialmediaToday – https://www.socialmediatoday.com/news/x-is-developing-a-grok-analysis-button-for-ai-fact-checks/716675/
