Example case:

* https://github.com/apache/airflow/issues/45904  - airflow
* https://github.com/apache/iceberg/issues/12034 - iceberg

Both issues are generic and useless and bring 0 value except noise.

Interesting thing is that many of those users, if you look at their history
- created. similar number of issues in iceberg and airflow about the same
time. So you might have more such issues now than you think.

J.




On Wed, Jan 22, 2025 at 9:41 AM Jarek Potiuk <ja...@potiuk.com> wrote:

> I have not counted all of them. there are quite a bit too many - and
> other people closed some of them as well. I got a very rudimentary check
> and applied "AI Spam" label to some of the issues
> https://github.com/apache/airflow/issues?q=is%3Aissue%20state%3Aclosed%20AI%20label%3A%22AI%20Spam%22.
> -> so we have had at least 25 such issues in the last 12 hours.
>
> > we also want to make sure that we don't accidentally close issues that
> don't come from a bot, but just a newcomer to the project.
>
> Those reports and patterns look very. very human-like - they are reported
> infrequently (per user) the description and text seem legitimate, but they
> are wordy and just reading and understanding that those are completely
> useless takes a lot of time. This is part of the problem, that it takes a
> lot of energy and time to determine if those are valid or not - and with
> such a rate, it's not sustainable just to analyze whether they are good or
> bad.
>
> J.
>
>
>
> On Wed, Jan 22, 2025 at 9:23 AM Fokko Driesprong <fo...@apache.org> wrote:
>
>> Hey Jarek,
>>
>> Thanks for bringing this to our attention. When you talk about flooding,
>> how many are we talking about? I see some suspicious issues (eg, here
>> <https://github.com/apache/iceberg/issues/12039>), but not many. I
>> hope this will come to a halt soon because it all additional work, and we
>> also want to make sure that we don't accidentally close issues that don't
>> come from a bot, but just a newcomer to the project.
>>
>> Kind regards,
>> Fokko
>>
>> Op wo 22 jan 2025 om 09:00 schreef Jarek Potiuk <ja...@potiuk.com>:
>>
>> > Hey Iceberg community, And Airflow community too.
>> >
>> > As of yesterday Airflow repo is literally flooded with a number of
>> issues
>> > that look almost good, except they are clearly AI generated and make no
>> > sense or repeat content from other issues. We noticed that the users who
>> > create a lot of the "spam AI" issues that are created in Airflow are
>> also
>> > creating similar issues for Iceberg.
>> >
>> > We got to the point that we are closing and reporting such issues to
>> > GitHub and we are blocking all such users without spending too much
>> time on
>> > it with messages similar to this:
>> >
>> > ```
>> > This looks totally AI-generated. useless issue report that brings no
>> value
>> > and makes no sense. We are generally blocking users that sends a lot of
>> > spam AI reports generated by bots.. as of yesterday so we will report
>> your
>> > account and block it unless:
>> >
>> > a) you explain how you generated reports
>> > b) prove you are human
>> > c) explain why you created the issue
>> > ```
>> >
>> > My guess is that some company released and is testing an "agentic AI"
>> that
>> > is "github-targeted" - where people can run the AI agents on their
>> behalf.
>> > It does not look like regular bot-spam.
>> > I think we should all generally crowd-source reporting it to Github -
>> and
>> > hopefully they will find a way to battle those without involving
>> > maintainers.
>> >
>> > I hope it will not last too long.
>> >
>> > J.
>> >
>> >
>> >
>> > ---------- Forwarded message ---------
>> > From: Jarek Potiuk <ja...@potiuk.com>
>> > Date: Wed, Jan 22, 2025 at 8:12 AM
>> > Subject: Re: Very strange (AI generated) issues
>> > To: <dev@airflow.apache.org>
>> >
>> >
>> > You can also report it directly from the issue (... at the top and
>> "report
>> > content")
>> >
>> > On Wed, Jan 22, 2025 at 7:46 AM Amogh Desai <amoghdesai....@gmail.com>
>> > wrote:
>> >
>> >> Elad, I just managed to report this user.
>> >>
>> >> This is how its done:
>> >>
>> >>
>> https://docs.github.com/en/communities/maintaining-your-safety-on-github/reporting-abuse-or-spam#reporting-a-user
>> >>
>> >> Thanks & Regards,
>> >> Amogh Desai
>> >>
>> >>
>> >> On Wed, Jan 22, 2025 at 12:05 PM Elad Kalif <elad...@apache.org>
>> wrote:
>> >>
>> >> > There are several reports from this user
>> >> >
>> >> > https://github.com/atharv9017
>> >> >
>> >> >
>> >> > I didnt find a way to report the user account to github.
>> >> >
>> >> > בתאריך יום ד׳, 22 בינו׳ 2025, 06:41, מאת Pavankumar Gopidesu ‏<
>> >> > gopidesupa...@gmail.com>:
>> >> >
>> >> > > Yes, still issues are coming.
>> >> > >
>> >> > > Regards,
>> >> > > Pavan
>> >> > >
>> >> > > On Wed, Jan 22, 2025 at 4:35 AM Amogh Desai <
>> amoghdesai....@gmail.com
>> >> >
>> >> > > wrote:
>> >> > >
>> >> > > > I saw a couple of such SPAM issues too.
>> >> > > >
>> >> > > > I also recall some SPAM comments on pull requests as well, so if
>> any
>> >> > > > contributor sees any such SPAM message,
>> >> > > > please report it on Slack so that we can delete it and report it.
>> >> > > >
>> >> > > > Thanks & Regards,
>> >> > > > Amogh Desai
>> >> > > >
>> >> > > >
>> >> > > > On Wed, Jan 22, 2025 at 8:45 AM Zhe You Liu <
>> zhu424....@gmail.com>
>> >> > > wrote:
>> >> > > >
>> >> > > > > I came across another strange issue:
>> >> > > > > https://github.com/apache/airflow/issues/45837. It appears to
>> be
>> >> a
>> >> > > > > copy-paste of https://github.com/apache/airflow/issues/45661
>> with
>> >> > just
>> >> > > > the
>> >> > > > > issue title changed.
>> >> > > > >
>> >> > > > > On Wed, Jan 22, 2025 at 6:50 AM Jarek Potiuk <ja...@potiuk.com
>> >
>> >> > wrote:
>> >> > > > >
>> >> > > > > > I even got to this stage:
>> >> > > > > >
>> >> > > > > > > We've received a few new tickets from your account
>> recently.
>> >> If
>> >> > > you'd
>> >> > > > > > like to add additional information you can add a comment to
>> an
>> >> > > existing
>> >> > > > > > ticket, or wait a few minutes before opening a new ticket.
>> >> > > > > >
>> >> > > > > > On Tue, Jan 21, 2025 at 11:49 PM Jarek Potiuk <
>> ja...@potiuk.com
>> >> >
>> >> > > > wrote:
>> >> > > > > >
>> >> > > > > > > There are few more that I still saw after sending it.
>> There is
>> >> > > > > something
>> >> > > > > > > going on bypassing GitHub filters.  I hope they will manage
>> >> to do
>> >> > > > > > something
>> >> > > > > > > about it
>> >> > > > > > >
>> >> > > > > > > Last one is https://github.com/apache/airflow/issues/45867
>> >> > > > > > >
>> >> > > > > > > On Tue, Jan 21, 2025 at 11:46 PM Vikram Koka
>> >> > > > > > <vik...@astronomer.io.invalid>
>> >> > > > > > > wrote:
>> >> > > > > > >
>> >> > > > > > >> Agreed.
>> >> > > > > > >>
>> >> > > > > > >> Thanks for flagging these Jarek!
>> >> > > > > > >>
>> >> > > > > > >>
>> >> > > > > > >> On Tue, Jan 21, 2025 at 2:34 PM Jarek Potiuk <
>> >> ja...@potiuk.com>
>> >> > > > > wrote:
>> >> > > > > > >>
>> >> > > > > > >> > Seems that we have a flood of AI generated feature
>> requests
>> >> > for
>> >> > > > > > Airflow,
>> >> > > > > > >> > The issues look somewhat legitimate, with somewhat
>> related
>> >> > > > content,
>> >> > > > > > but
>> >> > > > > > >> > they are wordy and make no sense when you read them.
>> Some
>> >> > > > examples:
>> >> > > > > > >> >
>> >> > > > > > >> > * https://github.com/apache/airflow/issues/45858
>> >> > > > > > >> > * https://github.com/apache/airflow/issues/45856
>> >> > > > > > >> > * https://github.com/apache/airflow/issues/45854
>> >> > > > > > >> >
>> >> > > > > > >> > All of them done by accounts with short history in GH
>> and
>> >> not
>> >> > > much
>> >> > > > > > >> activity
>> >> > > > > > >> > before
>> >> > > > > > >> >
>> >> > > > > > >> > There were quite a few more.
>> >> > > > > > >> >
>> >> > > > > > >> > I suggest we close such issues AND report authors to
>> >> GitHub -
>> >> > > > > > hopefully
>> >> > > > > > >> we
>> >> > > > > > >> > can help to battle the AI-generated traffic flood.
>> >> > > > > > >> >
>> >> > > > > > >> > J.
>> >> > > > > > >> >
>> >> > > > > > >>
>> >> > > > > > >
>> >> > > > > >
>> >> > > > >
>> >> > > >
>> >> > >
>> >> >
>> >>
>> >
>>
>

Reply via email to