I recently had the opportunity to chat with Max Spero and Bradley Emi, former Tesla and Google employees who are now the co-founders of Pangram — one of the most impressive AI detectors I’ve ever tested. I’ve used dozens of AI detection tools, but none have matched the accuracy of the one built by these two former Stanford roommates.
Although LLMs like ChatGPT and Gemini might help users write faster, Pangram is focused on what gets lost in the process — and how to restore trust in the written word.
With the ability to detect AI-generated and human-written content in over 20 languages, Pangram boasts over 99% accuracy for text in Spanish, French, Arabic and several others. Already used by large corporations and universities, Pangram gives users five free daily checks without a subscription.
Those with a subscription can take advantage of 600 AI scans a month (billed at $150/month), plus a browser extension, which is essentially like a black light over the internet to see what is human and what is AI-generated.
As someone who writes about AI for a living and put Pangram's AI detector to the test, I have to say it’s great for checking product reviews, evaluating website content for authenticity and so much more.
How Pangram works
By going to Pangram.com, users will see a box on the right to check for AI. Without logging in, users can paste their text or upload an entire document. From there, once the text is scanned for AI, users will see not only the results but understand the outcome. For example, even if something is written partially by a human, the tool will pull out the parts that were written by AI, clearly showing that it is a hybrid piece.
When AI is detected, the tool highlights words or phrases that look suspicious. At one point in my testing, after writing directly into the text box on the site, I had to go back to Max and ask why I was getting an “AI detected” result. This was odd to me especially since I was literally typing the words and not copying from Gemini or ChatGPT. Was it just because we ChatGPT is changing the way we speak? Turns out, I had used a marketing phrase that the tool considered most likely AI.
The browser is a game changer
I had a lot of fun testing the browser on Amazon. I checked everything from reviews to product descriptions and quickly discovered that a lot of text was written by AI. That’s not necessarily a bad thing. I think many are quick to judge, but in a world where AI is increasingly becoming part of our daily workflow, it’s something that we need to get used to.
Sellers might have an incredible product but not have the words to articulate what makes it great. Using AI writing tools can help them quickly get the words out in a more poetic way. But for me, AI generated reviews feel disingenuous if they have been completely written by AI.
By the way, Writer AI detector, didn’t detect a drop of AI text from the Amazon product description.
Pangram is detailed, but not perfect
As mentioned, there were times when a test appeared to be written by AI but it was actually written by a human. The team at Pangram acknowledged that they are still working on the few times this happened and appreciated the feedback.
For those eager to try this tool, especially teachers, professors or those in the corporate world, it’s important to acknowledge that some phrases, though human, have become overly used by AI, which might make them come across as wholly written by a bot. Pangram actually lists a few of them on the site. Does this mean we, as humans, need to stop using those words? I don’t think so. I still proudly use the em dash even though it’s become a marker for AI generated text.
Bottom line
After spending time with Pangram, I’m convinced this thoughtfully built platform is helping to solve how we can maintain trust in digital content. Pangram isn’t taking an alarmist stance, it offers transparency, giving context behind the content regardless of how the text was written (or generated).
I believe that level of nuance matters, especially as AI becomes more deeply woven into our everyday communication — from product listings and resumes to blog posts and school essays.
More from Tom's Guide
- ChatGPT Vision is useless for one thing — Google Lens nails it every time
- New study shows how people are using ChatGPT — and Google should be worried
- I use ChatGPT every day — but this one flaw still drives me crazy (and I can’t be the only one)
![HIDevolution [2025] ASUS ROG...](https://images.fie.futurecdn.net/products/60cbb9059616003fe7353df7bc0e56e02dd01b3c-100-80.jpg.webp "HIDevolution [2025] ASUS ROG...")
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
