OR confirmed that it is a false claim

x.com/openreviewne...

Did they report the issue via email? If it was X, I don't think they actively check that

Personally, I understand your point and want to agree, face the fact, and leverage the data for improving future peer-review systems

But as an organizer side for multiple venues, I am not sure

openreview.net/forum/user%7...

The attackers may have not taken their own actions seriously, but I am glad that the OR team took strong actions

I would give my kid HF Reachy Mini or programmable LEGO robot and distract him from such tracks 🫣

I wish I could come join 😭

Because it's a different topic

We are not seeing workshop papers (say NeuriPS workshop papers) as prestigious as such a business is advertising as "top AI conference papers"

However, highschool/undergraduate students and potentially their parents / school teachers would no think that way

I don't understand the question. Could you clarify/ rephrase?

I'm not going back to the Econ discussion

I am also concerned that the next level of unreasonable expectations on high schoolers and undergrads like "they have NeurIPS (workshop) papers" would be "and their papers were already cited X times, including those from such cool institutions"

Don't make it happen, please!

A bigger concern that I have is about accountability of such submissions / publications

Who's gonna take responsibility for the contents and wrongdoing if found?

Don't tell me that they'd throw high school / undergrad students under bus or "come on, those are just workshops"

It would be great if they could prepare check forms for detected reviews in OpenReview. It should not be that difficult

Then, PCs can monitor how many of the suspicious reviews were checked by ACs and share the report during the conference or a follow-up blog

Thank you for sharing the info

That indeed matches the language used in their blog post

It's understandable that they want to rely on ACs for this, but I have a concern that there may be many (not the majority, I'd say) ACs who don't do the job

Is the information detailed enough or just a list of suspicious review IDs ?

Not sure if it is true, but they mentioned that someone impersonated the authors and posted the public comments

If it's true, I guess PCs deleted the comment for the real authors?

As you can see in my original post, I am talking about LM API providers

I believe that uploading confidential information to their servers e.g. Google, OpenAI, Anthropic, etc is NOT ok

If the review platform offers a dedicated API and addresses the issue, it is a different story

I haven't seen any "good LLM reviews" yet and cannot agree

Besides, if we heavily rely on "frontier LLM" for reviews, just uploading papers to arXiv with the auto-generated review comment should be enough

Collected feedback may make it better, if the next AAAI PCs decide to take over the task and keep working on it

Yet, it still doesn't address this issue
> reviewers read papers that are not supposed to be disclosed to third-party