AI Generated Wayspot - Removal Request Denied

Wayspot Appeals
,
1

When submitting a Wayspot Appeal, make sure to include as much of the following information as possible:

  • Wayspot Title: Jackfish lake boat launch

  • Location (lat/lon): (53.482404,-114.236232)

  • City: Parkland County

  • Country: Canada

  • Screenshot of the Rejection Email (do not include your personal information):

    image
    image652×579 49.8 KB

  • Additional Information (if any): This is an AI generated wayspot on single family private property on Jackfish Lake in Alberta, Canada. On top of that, it is set about 70 meters away from the actual shore of the lake.

    image
    image899×394 63.7 KB

    image
    image723×531 106 KB

I believe this user has another wayspot under review currently that is also AI generated in roughly the same place. These can be easily verified at https://openai.com/verify for ChatGPT generated, using @Verify AI for Gemini generated.

2

image
image1199×852 162 KB

This is infuriating.

Do we need to pause Wayfarer until these can be caught if reviewers aren’t doing the job?

3

Pausing Wayfarer feels like the nuclear option. I don’t think we punish the entire community because some people are determined to game the system. However that being said - when something like this gets through review and then survives a removal request, we have a problem.
Two things -

  1. The community needs to slow down on reviews and make sure they are looking closer.
  2. The appeals reviewers could possibly include a detection tool to find the fakes.
4

I said this because it is my understanding that this is what Niantic Spatial had to do for Ingress. I have not kept up with Ingress or Niantic Spatial though.

5

Yeah, Recon had a lot of cheating issues. Far greater than anything Wayfarer has dealt with. This is not the Recon forum, so I won’t get into it here. I personally think there are tool out there to detect AI images used in a wayfarer nomination, but I don’t know what they use now.

I think maybe increasing the sensitivity might be one solve to the problem? Maybe?

6

I agree with @BlameJamal that I would not want the nuclear option :frowning: but do agree with @cyndiepooh that something needs to be done.

A few thoughts…

  • Introduction (or increases sensitivity) of AI checking so it doesn’t get to Reviewers in the first place.
  • An “AI” detected option in the Review flow.
  • An automatic “Send to Staff” if 1 review states AI / Fake especially from long term proven wayfarers.
  • Extra Warnings that any use of AI will not be tolerated on Submission screens.
  • Automatic Suspensions for any reviewer found to have used them.
7

I agree that having a means to explicitly flag wayspots under review (and existing wayspots) as AI generated/manipulated would be helpful. The generic ‘fake nomination’ option somehow seems like too much of a catch-all?

With that said, I’ve reported several very obvious and verifiable AI wayspots, and I included explanations and links to use to verify that they’re fake, but I’ve had to come to the forums to appeal two of them. Maybe I’m particularly sensitive to AI images and the uncanny valley, but the ones I’ve reported feel pretty blatant, even without using verification tools to prove they’re fake.

8

Thanks for the appeal, @flux780! We gave this a second look and have already retired the Wayspot in question.