In the event you pay for Twitter (sorry, X), you now have entry to Grok, the corporate’s AI bot. A part of that privilege consists of entry to a trending information feed within the Discover tab powered by the corporate’s synthetic intelligence. The one problem? It is rubbish.
Here is how Grok’s trending information feed appears to work: The bot pulls collectively the “high” posts regarding any given information story, then generates a abstract of the information from these posts. Easy sufficient, and one thing that we have seen generative AI do earlier than. However earlier than you fireplace your human writers and editors and put Grok in control of the information, you may need to see how precisely it is reporting works.
Grok seems to be aggregating joke tweets and spitting out an AI-generated reply as real information. You possibly can see that from this submit shared by X person BrettRedacted following the earthquake that shook a lot of the New York Metropolis metro space. The bot generated the headline: “Adams vs. Earthquake: 50,000 Cops in Subway Showdown,” then stories how New York Metropolis Mayor Eric Adams is deploying the NYPD to “forestall additional earthquakes,” contemplating utilizing “robo cops,” and has ordered “each cop within the metropolis” to “shoot the rattling earthquake earlier than it strikes once more.”
Positive, we dwell in bizarre occasions, however no affordable individual would ever consider that Grok’s abstract of the information was correct. In the event you by some means could not determine for your self, you possibly can check out the highest posts fueling that information abstract, which on this case, are all joke tweets in regards to the mayor’s response to the earthquake.
This can be a unhappy however humorous instance of the state of X in 2024. Previous variations of the location would have been the place to be to observe each official updates on breaking information just like the NYC earthquake, and snigger on the jokes in regards to the scenario. Now, the location treats the jokes as information. You get what you pay for, I suppose.
It does not take a lot foresight to think about how this case goes from unhappy however humorous to easily harmful. What occurs when Grok decides to “report” on one thing that appears at first look official, however relies on the rampant misinformation spreading throughout the location? An X person might verify into the Discover web page to see that their metropolis is being bombed, or {that a} candidate of their election did one thing unlawful, when neither “story” is definitely true.
Whereas we won’t cease the location from pushing this crap, we are able to all collectively comply with not deal with Grok, or any generate AI for that matter, as a official information supply, or an correct summarizer of the day’s high tales.
