Elon Musk has always expressed concern about AI’s risks. He doesn’t seem to be too concerned, though, about X’s own defective AI chatbot, Grok, which recently produced and disseminated erroneous news articles.
It has been alleged that Musk is so indifferent to Grok’s problems that he is now trying to push it even farther throughout X. He has given the social media platform’s programmers the task of incorporating the AI chatbot directly into its tweet composer.
In a recent story, Kylie Robison of Fortune reports that Musk wants Grok to be accessible for X Premium members directly in the tweet box, allowing Grok to compose tweets on their behalf. According to the source for Fortune, Musk “wants people to sound smarter” and he thinks Grok would help achieve that.
The issue with Grok
Grok writing user posts raises a number of serious problems. As Fortune’s source claims, X staff are aware of this and have attempted to delay the release of this functionality.
For starters, the spam and bot issue has only become worse since Musk took over the website that was once known as Twitter. Furthermore, a large number of these spammers and bots are obviously driven by AI chatbots and writing tools. One can only assume that this issue will worsen if X incorporates such a feature directly into the platform through Grok.
Furthermore, the report claims that Musk’s x.AI API, which drives Grok, is too slow, making it difficult for X engineers to utilize.
Additionally, as was already said, there is the whole problem where Grok has been known to fabricate articles and pass them off as true. These stories are then encouraged by X’s newly added Explore and trending topics sections.
Ed Zitron of the Better Offline Podcast also brought out the fact that X’s Grok AI chatbot has already been educated using information from users’ tweets as another possible issue. Grok will really get worse if users utilize it to create posts, and then Grok uses that content to train itself. Researchers have discovered that training an AI model with AI-generated content causes flaws in the model.
Musk had asked months ago for Grok to be included in the tweet composer for customers of X Premium subscriptions costing at least $8 per month. Though it hasn’t happened yet, this will probably wind up being done eventually unless X’s engineers can persuade him that this is a poor idea. Prepare yourselves, then, for an X that contains more bots and AI-generated spam than ever before.