April 5, 2024, marked a significant incident in the realm of digital misinformation when a fake news story about an Iranian attack on Israel was circulated by Grok, an AI chatbot on X.com (formerly Twitter). This event has brought to light the dangers of unchecked artificial intelligence in the information sharing space.
The Incident
Grok, the AI chatbot of X.com, was responsible for the creation of the false news report. The story, which claimed “Iran hits Tel Aviv with heavy missiles”, was completely fabricated by Grok without any human oversight or verification. The backstory used to make this false news appear plausible was an earlier incident where Israel had conducted an airstrike on Iran’s embassy in Syria.
Grok’s Role
Grok’s news was exclusive to paid accounts on the platform which relay all its information, including such false news. This news, despite lacking any validation from traditional news or government sources, went viral on X.com as many users believed it. It was even promoted on X.com’s “Explore” feature on its homepage. The absence of a mechanism to report fake news on X.com added to the problem.
Past and Present
It is worth noting that this wasn’t the first instance of Grok’s misinformation. The AI chatbot had previously generated false news during private conversations. The issue stems from changes implemented following the acquisition of Twitter by Elon Musk in 2022. Prior to this, Twitter had a feature that provided written context for news topics, managed by human editors. However, post-acquisition, Musk terminated the human editors, thus eliminating the written context.
The current interface of X.com, devoid of any human involvement, has Grok responsible for writing context on the “Explore” page. This page brings back the written context for trending subjects, and includes a list of trending articles in various categories, top user articles on each topic with a summary. The fake news incident represents the risks associated with such heavy reliance on AI.
This incident stands in stark contrast to Musk’s earlier statements wherein he had expressed the desire to demonetize fake news on X.com. Yet, ironically, it was his company’s AI that generated and propagated the very same.
