Experts warn against using Grok AI for information verification

One of the most common responses seen lately under posts on X (formerly known as Twitter) is the question, "Is this true, Grok?" Experts, noting that Grok is an AI chatbot that can also produce false information, express concerns that using the system as a verification tool or fact-checking center could lead to significant issues.

After Elon Musk purchased Twitter and renamed it X, one of the major updates was the introduction of Grok, an AI-powered assistant. This system, in addition to its various capabilities, became known for answering user queries. Initially, Grok was used to identify the song playing in a video or which movie a certain scene belonged to, but it has now taken on a broader role as a "verification center" in a wide range of fields, including politics, economics, military issues, sports, and entertainment.

The number and diversity of questions directed at Grok have increased significantly. While many answers are correct, some mistakes are casting doubt on Grok's reliability. The main question here is whether Grok is truly designed for such a mission. Additionally, another concern is how possible incorrect answers may impact society.

Ersin Çahmutoğlu, a cybersecurity expert from ADEO, is one of the prominent figures who closely follows developments in the cyber world and similar fields in Türkiye. Before addressing the issue of Grok's answers, Çahmutoğlu provides a brief outline for those unfamiliar with the topic. He defines Grok as an "AI-powered chatbot" and explains that it is one of the innovations introduced after Elon Musk's acquisition of Twitter and its rebranding to X.

Çahmutoğlu shares that Grok was not very popular in its early stages, but after the addition of features like profile visualization, news analysis, and others, the number of users increased significantly worldwide. He also notes that, in line with global interest, Grok has begun to be seriously used in Türkiye as well. When examining posts on X, it is evident that one of the most frequent responses in the comment section is the message "Is this true, Grok?"

This is an example of Turkish users consulting Grok for verification. However, the main issue here is not asking questions. Çahmutoğlu says, "Users who tag the X account to ask questions tend to accept the AI's answers as true. This is where the danger lies."

He emphasizes that Grok, as an AI robot, cannot know everything, pointing out that "there are errors in some responses. But in Türkiye, this issue has almost turned into a trend. People even seek Grok's verification for current political issues or breaking news. That's not Grok's purpose."

This is where the issue of using Grok as a verification tool comes into focus. Çahmutoğlu warns that providing wrong answers to questions related to social events could lead to serious problems. He stresses that Grok should not be used as a fact-checking line, and adds:

"As we mentioned earlier, Grok is an AI chatbot… Yes, in some cases, it can be useful for analyzing certain claims or developments that have become widespread. However, this does not reflect the reality that it knows everything correctly.

If, in the future, Grok gets a feature for fact-checking, perhaps more accurate posts could emerge. But even then, it would still be necessary to verify information from different sources, because the multi-comparison approach on social media always provides more accurate results.

At this point, users must absolutely not forget one thing… Especially information or news that could affect a large portion of the public or trigger certain groups, asking Grok and accepting its answer as absolute truth poses serious risks.

Instead of asking Grok about a breaking news item on social media, it would be more appropriate to check more reliable media outlets or primary sources. To ensure Grok is used properly in Türkiye, there needs to be a growing awareness on social media."