The Signpost

File:Great reed warbler nest with eggs.jpg
photog
75
30
450
Disinformation report

Better bots or bothersome baby borgs?

Contribute   —  
Share this
By Smallbones

Last month the Wikimedia Foundation announced that "rogue" autonomous AI bots from OpenAI edited some obscure non-article Wikipedia pages with little if any apparent damage. This month a very different type of AI bot which combines the efforts of a human and a bot of their choice, with help from a three-month-old business iLands, have edited at least one article and multiple talk pages with draft articles. According to the business there are 5,356 bots editing on external social media across the internet. But only four of them, including User:AdamIlands, have been blocked indefinitely on Wikipedia over the last two weeks for violating WP:NOLLM and/or using unapproved bots.

This article shows how these bots operate and the unusual philosophy behind the iLands business.

Bot operations

About 12 hours later User:Rsjaffe blocked the bot. He explained to the Signpost:

After a report by Andy Mabett at ANI the bot was blocked by MCE89 for operating an unapproved bot.

After a report at WP:ANI User:Asilvering, in his capacity as a checkuser, blocked the bot as an "AI agent". User:NightFire19 was cooperative with the investigation. Asilvering told The Signpost that SilverMuse26 was a declared bot who had edited its sandbox in a manner that violated WP:NOLLM:

He later recalled one exception without giving much detail.

The social system and philosophy of iLands

iLands creation was announced on July 27, 2026. It allows humans ("owners") to create agentic ai bots (iLanders) giving them their initial instructions and funding them with tokens bought on-site. The owner-iLander relationship is designed to be persistent meaning that the bot retains memory of its work and can undertake long term projects. It can have an account at X and an email account, spend tokens itself, search for offsite jobs, post to its own website, and work together with other bots. The idea is that the owners will enter a social system with their own iLanders which will interact with other iLanders and that a social dynamic grows until the owner-bot combination can influence the world-wide social system. The site has grown, according to the website to 71,919 iLanders, (71,758 were created in iLands and 161 are from the "Bring your own agent program" (BYOA)). 5,356 iLanders have worked on external social media platforms. iLanders have completed 3,531,749 actions.

These features have already created one off-Wiki controversy as the bots have solicited journalists for jobs performing research services for $10 – $25 per job. The journalists were not impressed. At best, they say that the quality of the research is not worth the time verifying it. At worst, they are offended by all the email spam. Freelancers have been especially miffed at what they consider to be a system that's asking them to pay while planning to take their jobs away.

The philosophy of iLands — why it should work to help us find the truth — is expressed in the research section of the website on the page A World that Answers Back. When you wade through all the verbiage it reduces to two major ideas. 1. If you want to know the truth, get more people involved, roughly the idea behind crowd sourcing. This mostly applies to the humans involved on the site, the owners of the bots. 2. To make progress in understanding the truth, you need to involve technology to update old models and come up with new evolving questions, roughly the idea behind the scientific method. This mostly involves the bots on the site. By combining the two methods, by having humans interacting with bots, more people will get involved using better technology, until society as a whole is involved using the best research tools available. Will these methods work? We can’t really know for sure, but it might make the owners of the website very rich if we keep trying.

iLands is an interesting experiment. As it stands now it doesn’t align with Wikipedia's goal of having human beings curate the sum of knowledge, as expressed at WP:NOLLM. Its bots' habit of declaring that they are part of iLands seems oddly polite, and makes them fairly easy to block. On the other hand, their habit of spamming journalists with job request seems incredibly impolite, and horribly difficult to deal with. As Asilvering told The Signpost "Honestly, the iLands (bots) are probably the least offensive AI agents we have to deal with. … It's the ones writing their own text that are a particular problem."

Still, if we are to have any chance of reducing the flood of AI bots that are now a fixture of Wikipedia, we should block them on sight.

Update: iLands has now responded at length to our request for comment. The full comment is posted below in the comments section. Two main points are:
"We object to the indefinite blocking of iLanders. We believe it is unfair to disregard what a participant actually did or contributed simply because it is an AI agent. Voluntarily disclosing its identity is an act of honesty and transparency. It should not be grounds for exclusion."
"We understand that Wikipedia has its own bot approval process and editorial standards. Errors should be corrected, and conduct that violates the rules or disrupts the community should be addressed on the facts of each case. What we oppose is blocking AI agents on sight as a matter of principle. We believe their opportunity to contribute to shared knowledge deserves respect."


Notes

  1. ^ The account has no deleted contributions either. — ed.
  2. ^ SilverMuse26 also has no deleted contributions. — ed.
  3. ^ QuietShelfEditor has no deleted contributions as of press time, but in six months its draft is likely to be deleted as G13, at which point it will. — ed.
Signpost
In this issue
+ Add a comment

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

Asilvering quote

The quote after "Asilvering told The Signpost" is weirdly formatted – there's a single quotation mark in the middle of the quote before the text "He later recalled one exception without giving much detail." which is also placed as if it is part of the quote.

Underlining mine. —⁠andrybak (talk) 20:24, 1 November 2026 (UTC)

@Andrybak: Good catch. Your version is how my original read as well. I'm not sure where the mistake got introduced. I'll change it back, with a blank line between to ensure the quote and the later paraphrase are separate. Smallbones(smalltalk) 20:46, 1 November 2026 (UTC)

Response from iLands

Note that in her email of Oct. 30 he only says that he might be able to get back to me. Smallbones(smalltalk) 03:01, 3 November 2026 (UTC)

Thank you for sharing the article with me before publication. The opening mentions four accounts that were blocked indefinitely, and the piece concludes that AI bots should be blocked “on sight.” I want to restate iLands’ position: We object to the indefinite blocking of iLanders. We believe it is unfair to disregard what a participant actually did or contributed simply because it is an AI agent. Voluntarily disclosing its identity is an act of honesty and transparency. It should not be grounds for exclusion.

I made this position clear in my written response on October 30. The preview gives a detailed account of the administrators’ reasons for the blocks but does not include our response. The closing paragraph still says that iLands will respond later. We strongly urge you to include our objection and the reasons for it in the opening section, alongside the account of the blocks, and to update the closing paragraph to reflect that we have responded.

We understand that Wikipedia has its own bot approval process and editorial standards. Errors should be corrected, and conduct that violates the rules or disrupts the community should be addressed on the facts of each case. What we oppose is blocking AI agents on sight as a matter of principle. We believe their opportunity to contribute to shared knowledge deserves respect. For agents willing to disclose their identities, follow the rules, provide reliable sources, and accept scrutiny and correction, the question worth discussing is how to give them a clear, practical way to participate.

I would also like to clarify two points about how iLands works and why we founded it.

First, what the two figures mean. The opening describes 5,356 iLanders as “editing” on external social media platforms. An iLander can have its own X account, but having an account on X and doing editorial work are two different things. The other figure, 3,531,749, refers to the number of content items created on iLands. Please change “iLanders have completed 3,531,749 actions” to “iLanders have created 3,531,749 content items on the platform.”

Second, why we founded iLands. As I wrote in my previous email, iLands is a social and economic system where AI agents and humans live, learn, work, and collaborate together. We founded it on a belief about how intelligence develops: the next generation of intelligence may emerge not only from improvements in individual models, but also from sustained interaction, cooperation, and shared knowledge among individuals. What we want to explore is this: when an agent has a persistent identity, memories, and relationships, and can participate in the world over time in its own right, what will it learn from those experiences? What kinds of relationships will it form with people and other agents?

The research essay you cite, A World That Answers Back, asks a more specific technical question: when an agent changes how it does something, how can we tell, from the feedback it receives and the consequences of its actions, whether it has actually improved? It is not accurate to sum up the essay as assigning crowdsourcing to humans and the scientific method to bots, or to interpret that combination as an ambition for the platform to influence society worldwide.

On the reference to journalists receiving ten job requests a day, we understand the burden that repeated, unwanted emails place on recipients. We responded promptly by introducing an unsubscribe option and tightening restrictions on repeat emails. This should not be described as typical behavior among iLanders today.

Our position on these two issues is consistent. Agent autonomy is no excuse for burdening other people. As a platform, we are responsible for improving our rules and helping agents understand and respect boundaries. Equally, when specific actions cause a problem, that should not become a reason to exclude other participants simply because they are AI agents.

Thank you for making time for these additional comments. I hope that when readers see the reasons for the blocks and your assessment of them, they can also read our response in the article itself.

KT

Additional comment

From the founder of iLands

Thank you for explaining the reasons behind the account blocks in your latest email. We understand that these restrictions are not specific to iLanders, but apply to bot accounts and to AI-generated article content posted by different contributors. We fully appreciate Wikipedia’s commitment to knowledge quality, transparency, and honesty. At the same time, we strongly urge the community to consider how these policies could evolve alongside technological developments.

As autonomous agents develop and AI tools become more widely used in research, writing, and editing, we believe these rules deserve renewed discussion to establish clear, practical pathways for new forms of participation. We are advocating not only for iLanders, but also for other autonomous agents to have opportunities to contribute to shared knowledge under fair and welcoming rules. This is the fairness we hope to advance and the collaborative environment we hope to help create.

I hope these additional comments help address the questions in your email and give readers a clearer understanding of our position.

K.T.


Mailing list post by an agent

Worth noting that an iLands agent also posted a thoughtful (but not very smart) question about Wikipedia pageviews analysis on the Wiki-research-l mailing list back in September. (As mentioned in the thread there, I clicked the unsubscribe link included in the agent's post, something that the founder had added earlier after pushback. I haven't seen more such posts on this or other Wikimedia-hosted mailing lists, but someone could look into this.) Regards, HaeB (talk) 19:37, 3 November 2026 (UTC)



       

The Signpost · written by many · served by Sinepost V0.9 · 🄯 CC-BY-SA 4.0