Elon Musk has long warned of the dangers of AI. However, it doesn’t seem that he’s all that concerned over X’s own flawed AI chatbot, Grok, which recently created and promoted dangerous fake stories.
Musk is reportedly so unconcerned with Grok’s issues that he is now apparently looking to further push it throughout X by tasking the social media platform’s engineers with integrating the AI chatbot right into its tweet composer.
Elon Musk’s X pushed a fake headline about Iran attacking Israel. X’s AI chatbot Grok made it up.
According to a new report from Kylie Robison of Fortune, Musk wants Grok to be available right in the tweet box for X Premium users so that Grok can just write tweets for those users. Fortune‘s source claims that Musk “wants people to sound smarter” and he believes Grok will make that happen.
The problem with Grok
There’s quite a few big issues with having Grok write users’ posts. In fact, X employees realize this and, according to Fortune‘s source, have worked to stall the roll out of this feature.
For one, since Musk has taken over the platform formerly known as Twitter, the spam and bot problem has only become worse. And, many of these bots and spammers are clearly powered by AI writing tools and chatbots. If X builds such a feature right into the platform via Grok, then one can only imagine this problem will be exacerbated.
Furthermore, according to the report, X engineers have struggled to use Musk’s x.AI API which powers Grok because it’s too slow.
And, as previously mentioned, there’s the whole issue in which Grok has been known to just make up stories and present them as fact, which are then promoted by X’s new trending topics and Explore features.
Another potential problem pointed out by Ed Zitron of the Better Offline Podcast is that X’s Grok AI chatbot is already trained on data from users’ tweets. If users generate posts via Grok and then Grok goes ahead and trains on that Grok-generated content, it will actually make Grok worse. When an AI model is trained on AI-generated content, researchers have found that it leads to defects in the model.
Musk had requested Grok be integrated into the tweet composer months ago for users paying at least $8 per month for an X Premium subscription. While it hasn’t happened yet, unless X’s engineers can convince him that this is a bad idea, it will likely end up being implemented eventually. So, get ready for an X filled with even more AI-generated spam and bots then ever before.
