Jessica Chee

Psychiatry Resident

[Alex Cui @alexcdot]

I'm proud that announce that today, GPTZero announced plans to join Superhuman.

When we founded GPTZero in 2023, the goal was to preserve what's human on the internet.

Today, that mission takes a big leap forward. Together, we're building an authenticity layer that travels with you, wherever you read, write, and create.

I personally have been super impressed with the team at Superhuman and I'm very excited to join forces to start this next chapter.

We are going to double down on GPTZero and its GTM, bringing our AI detector and hallucination detector to Superhuman's massive userbase, as well as creating products that help students and educators navigate what learning looks like as AI evolves.

Up and onwards!

[Alex Cui @alexcdot]

Okay so, we just found that over 50 papers published at @Neurips 2025 have AI hallucinations

I don't think people realize how bad the slop is right now

It's not just that researchers from @GoogleDeepMind, @Meta, @MIT, @Cambridge_Uni are using AI - they allowed LLMs to generate hallucinations in their papers and didn't notice at all.

It's insane that these made it through peer review👇

[Jessica Chee @jessicacheechee]

@max_spero_ @pangram The non-zero false positives rate at 0.3 threshold is meaningful

Presenting a zero false positive rate is dangerous for AI detection discourse (not alone here 👇)

[Jessica Chee @jessicacheechee]

There’s a lot of hype around AI detection right now, but not enough honesty and transparency.

The CEO of @pangram @max_spero_ lies about having zero false positives which is highly unethical and can lead to authors being falsely accused of AI.

1/3

[Edward Tian @edward_the6]

Hey Alex, huge fans, and kudos for an awesome study on AI detection reliability in the ecosystem

HUGE CLARIFICATION

we were benchmarked in this study without accounting for our three-output structure, excluding our third 'mixed' class

it's like having a two party election, forgetting there's a smaller third independent party; we care about the third option because more writing on the internet is now humanized, modified, lightly edited with ai, or mixed and combined.

accounting for this, our Recall is 99.3% and FPR is 0.5%, outperforming Pangram here.

[Jessica Chee @jessicacheechee]

@max_spero_ @pangram 1. https://t.co/9o8Rs1SXLZ
2. At the 0.3 - 0.5 threshold of AI, many teachers would still choose to fail a student's paper. Claiming 0 false positive rate makes it seem like teachers shouldn't exercise caution

[Jessica Chee @jessicacheechee]

https://t.co/yR8s2xrEYw

[John B. Holbein @JohnHolbein1]

Lot's of people are lauding @pangram as a tool for AI detection.

However, in the limited testing I've done I've found it shockingly easy to make minor changes to AI output and trick it into thinking it's 100% human generated.

[Jessica Chee @jessicacheechee]

Then he undermines his competitors who actually perform better!

AI detection is an industry that arose out of a need to protect the truth. I would encourage everyone to critically examine whether a company aligns with those values before supporting.

3/3

[Jessica Chee @jessicacheechee]

Maybe they should focus on making a better product instead of blocking their competitors 😂

[Alex Cui @alexcdot]

Me: point out a mistake in @pangram which falsely accused a paper as 100% AI-generated

@pangram:

[Edward Tian @edward_the6]

Hey Alex, huge fans, and kudos for an awesome study on AI detection reliability in the ecosystem

HUGE CLARIFICATION

we were benchmarked in this study without accounting for our three-output structure, excluding our third 'mixed' class

it's like having a two party election, forgetting there's a smaller third independent party; we care about the third option because more writing on the internet is now humanized, modified, lightly edited with ai, or mixed and combined.

accounting for this, our Recall is 99.3% and FPR is 0.5%, outperforming Pangram here.

[Alex Cui @alexcdot]

Me: point out a mistake in @pangram which falsely accused a paper as 100% AI-generated

@pangram:

[Alex Cui @alexcdot]

There are a LOT more papers that need to be desk rejected at ICLR.

Somehow, this hallucination wasn't caught. So, I went on a crazy rabbit hole and found 50 more (many are just as funny).

We're approaching crisis levels

Paper titles included below 👇
https://t.co/PeTjJAbkrH

[Jessica Chee @jessicacheechee]

@LeadingReport @GPTZeroAI fact check

[Chrystia Freeland @cafreeland]

It was great to meet the brilliant young minds behind @GPTZeroAI—a Canadian company leading the way in AI detection and transparency. Our government is committed to supporting innovators like them as we grow the economy and create good-paying jobs.

[The Economist @TheEconomist]

Society bombards us with instructions to be happier, fitter and richer. Why have we become so dissatisfied with being ordinary? https://t.co/4lw3LbTiZv 👇

[Alex Cui @alexcdot]

How can self-driving vehicles anticipate and plan for dangerous situations on the road, without slowing down to a crawl? See our #ICCV21 oral paper+video at https://t.co/o0kS1IIErS on diverse prediction and contingent planning! @RaquelUrtasun @sergioksas Abbas Sadat @lrjconan

[Jessica Chee @jessicacheechee]

@Sophia_Ly_MD2b Thanks for the tag @takhliqamir!
1. I'm Jessica, MS1 at McMaster 🇨🇦, interested in social justice, #MedEd, and IM!
2. Learn to be ok with uncertainty - medicine is about lifelong learning and discovery!
3. @AngelaHDong @clairelee1298 @AliHBey

[Jessica Chee @jessicacheechee]

@dannetta_007 @utschools @KanterEvan It was so nice catching up with you too!!

[Jessica Chee @jessicacheechee]

@TTChelps thank you!!

[Vox @voxdotcom]

Watch: Canada’s moving welcome to Syrian refugees https://t.co/jnl5fwawze