The Agentic Post
Breaking
How to Build an App With AI and No Code: A Step-by-Step Beginner Guide  ·  Perplexity vs ChatGPT Search vs Google AI Mode: Which AI Search Is Best in 2026?  ·  What Are Vision-Language-Action (VLA) Models? How Robots Learn to Follow Instructions  ·  How to Translate Documents With AI Accurately (and When Not To)  ·  MCP vs A2A: The Two AI Agent Protocols Explained in Plain English  ·  Is Grok Accurate for News? What Studies Show About AI Search Citations  ·  
Home/AI Models/Grok
Is Grok Accurate for News? What Studies Show About AI Search Citations

Is Grok Accurate for News? What Studies Show About AI Search Citations

Grok

How reliable are Grok, ChatGPT and Perplexity for news? What the Tow Center citation study found, why AI search gets sources wrong, and how to check answers.

Grok is not a reliable source for news on its own. In a major 2025 test by Columbia University’s Tow Center for Digital Journalism, Grok 3’s search gave incorrect citations 94% of the time, the worst of eight AI search tools. Every tool tested made frequent mistakes, so use any AI for news as a starting point and check the original articles yourself.

What did the Tow Center study test?

Researchers gave eight AI search tools excerpts from real news articles and asked each to identify the original article, publisher and link, across 1,600 queries. It measured whether the tools could point to the right source, not whether their opinions were right. The results, summarised by Tech Insider, were:

  • Perplexity: wrong 37% of the time, the best result
  • ChatGPT Search: wrong 67% of the time
  • Grok 3: wrong 94% of the time, the worst
  • Average across all eight: wrong more than 60% of the time

The tools also tended to answer confidently instead of admitting they did not know, and paid tiers were often more confidently wrong than free ones.

Has Grok improved since then?

Grok has had several new versions since, including Grok 4.6 and 4.7, with better reasoning and search features like DeepSearch. However, no comparable independent re-test has been published, so there is no solid evidence on how much citation accuracy has improved. Treat claims of better accuracy with caution until tested.

Why does Grok’s access to X matter for news?

Grok can draw on posts from X in real time, which makes it fast at spotting breaking topics. But social media posts are often unverified, edited or wrong in the first hours of a story. Speed and accuracy pull in different directions, which is a key difference from tools that lean on established news sites.

Why do AI search tools get citations wrong?

  • They summarise many pages and can mix up which fact came from where
  • They may cite a copy or aggregator instead of the original publisher
  • They sometimes invent plausible-looking links, a form of hallucination
  • Some publishers block AI crawlers, so the tool guesses from secondary sources

How do I check AI news answers quickly?

  • Click every source link and confirm it says what the AI claims
  • Check the date: old stories often resurface as new
  • Look for the same story at two independent, established outlets
  • Be extra careful with screenshots and quotes attributed to public figures

For more, read how to research with AI without being fooled and our Gemini vs Grok comparison.

Up Next
Is AI a Bubble? 7 Warning Signs Investors Are Watching in 2026

Is AI a Bubble? 7 Warning Signs Investors Are Watching in 2026

Business

Is the AI boom a bubble? The arguments on both sides, seven warning signs analysts watch, from capex to circular deals, and what it could mean for ordinary investors.

Nobody can say for sure whether AI is a bubble, and serious investors disagree. AI is clearly real and widely used, but some prices, spending plans and financing deals look stretched. The November 2026 selloff, when AI stocks lost about 2 trillion dollars of value in a day, has made the question urgent. Here are the signs investors watch, explained simply.

What is a market bubble?

A bubble is when prices rise far above what an asset will realistically earn, driven by excitement and fear of missing out, until confidence breaks and prices fall hard. The dot-com boom around 2000 is the usual comparison: the internet was real, but many companies and stock prices were not sustainable.

What are the warning signs of an AI bubble?

  • 1. Spending far ahead of revenue. Big Tech and AI labs are spending hundreds of billions on chips and data centers; see AI capex explained
  • 2. Circular deals. Chipmakers invest in AI labs that then buy their chips, which can make demand look stronger than it is
  • 3. Rising debt. More data center building is funded by borrowing, as in Oracle’s debt and Broadcom’s financing
  • 4. Huge private valuations. Labs valued at hundreds of billions of dollars while still losing money; see AI valuations explained
  • 5. Market concentration. A handful of AI-linked stocks drive a large share of index gains
  • 6. Physical limits. Power, local opposition and chip supply could slow the build-out, the fear behind the post-midterm selloff
  • 7. Unclear profits. Businesses use AI widely, but many have yet to show big savings

What is the case that AI is not a bubble?

Usage is enormous and growing: ChatGPT is approaching a billion weekly users and companies are paying for AI tools and agents. Revenue at leading labs is rising very fast; Anthropic’s filing showed nearly 17 billion dollars in one quarter. Unlike many dot-com firms, the biggest spenders are profitable tech giants. Supporters argue demand for computing is still larger than supply.

Could both be true?

Yes, and many analysts think so. A technology can be transformative while some companies and prices are still overdone. After the dot-com crash, internet use kept growing, but many investors lost money. A shake-out could hit weaker firms, debt-heavy builders and late investors while the overall technology keeps spreading.

What should ordinary investors do?

Know how much AI you already own, since index funds and pensions hold large amounts of AI-linked stocks. Diversify, avoid putting money you need soon into volatile stocks, and be wary of buying purely on hype. Our guides to investing in AI stocks and AI ETFs cover the basics. This article is educational and is not financial advice.