Talk:Artificial intelligence

Scope?

This article is wordy and I'm not sure how it's directly relevant to consumer rights. Scraping the internet and data collection may be seen as unethical, but they're pretty run-of-the-mill at this point. Beanie Bo (talk) 01:56, 29 August 2025 (UTC)

I agree that the article needs significant revision. It has more detail than needed on some areas (e.g. web scraping), and totally misses other important areas.

I see AI more as a theme/background article. AI is so pervasive now, and affects people in so many ways, that I think it makes sense to have at least one article on it.

Things that I think such an article should cover include:

• Data centers - environmental impacts, community impacts, energy demand and subsidy by electricity and water rate payers, and how many of these agreements are made in secret, even in nominally democratic/open governmental systems. In the US data centers are often located in marginalized communities, where people are not as organized to protect their community . (This is not exclusively an AI thing might be worth a separate article about data centers in general, covering crypto mining operations, etc.)
• Inaccuracy and inappropriate use of LLM. "Hallucinations" People not understanding what an LLM is and assuming they are more capable than they are. LLM make a poor substitute for human written product reviews. (Inaccurate, praises whatever the user wants - even products that don't exist.)
• Control of information - Use of LLM in place of search is decimating independent information sources (taking away advertising revenue, taking away views).
• Intellectual property - piracy in training data (using stolen data), use of output.
• Privacy and security - data poisoning, ease of subverting guardrails, gathering data for training, revealing prompts, law enforcement review of chatbot prompts and outputs, etc.
• Concerns about possible effects on users - AI psychosis, etc.
• Labor concerns - conditions of labelers/piece workers.
• Liability - LLM are often inaccurate, what happens when the AI harms people (libel, suicide, etc.)

I have sources for a bunch of this, will be adding them to the article talk page as time permits. Drakeula (talk) 18:27, 4 September 2025 (UTC)

You're completely right. My mistake. This article does need significant reworking to maintain relevance, and a lot of the technical details should be simplified to maintain the wiki's voice and tone. But it's pretty relevant overall, so with time, it could fit better. Beanie Bo (talk) 19:02, 4 September 2025 (UTC)

posted this on the moderator's noticeboard, but posting it again here:

I'd caution here that I think quite a few of the practices listed probably wouldn't be within scope.

Certainly the following:

- Labour concerns

- Intellectual property

- Control of information/search blocking

- Environmental/social impact of data centres

Feel like they're out-of-scope as they concern relationships not relevant to the wiki, between businesses and other businesses/creatives, businesses and their employees/workers, as well as between businesses and the wider environment. To prevent scope creep, we want to keep the wiki focused on the consumer-rights issues.

And these ones feel like 'edge cases' for relevancy - I'd appreciate some wider input:

- Possible effects and harm on users from improper function (I'd argue that in a lot of cases there's not much to be done on this front, but I think if insufficient steps are taken to warn and safeguard users, then they could be mentioned. Certainly things like character.ai and similar do feel very exploitative, but I'm not sure I'd bundle the normal assistants under the same umbrella there)

- Liability (I'd say this can be relevant, but the emphasis should be placed on situations where people create systems using AI that take decisions that really shouldn't be left to AI, and harm consumers that way. This is always going to be a fuzzy line, and I'd expect extensive discussion over it - it feels analogous to the question of 'at what point does someone getting injured by their own chainsaw go from being manufacturer negligence, to user error?')

More broadly, I think that 'AI' probably isn't the best title for an article, as it's such a wide field. AI technically includes almost anything done by a computer. If we go by dictionary definitions, the computer opponents in old strategy games would count as 'AI'. LLMs, Generative image/video models, and traditional ML stuff like image recognition are all distinct enough, and are related to different issues, that it feels like they'd be better separated into their own articles, rather than bundled. Keith (talk) 16:04, 8 September 2025 (UTC)

Since AI is so pervasive as a marketing term, I think it appropriate to have an article on it. If nothing else, to give an easy to find roadmap of other related articles that cover aspects. For example, I would not expect the layman to know what generative AI, or LLMs are - even if much of the content of concern here winds up under such sub-articles.

I agree that subdivisiion probably makes sense. However, at this point I don't have a clear idea of what bits to split off. Drakeula (talk) 08:51, 13 September 2025 (UTC)

I am unclear why those particular items seem out of scope. Perhaps you could give more detail? What scoping rule do they violate? Is it because they are primarily "old" consumer concerns, rather than "new" ones? Drakeula (talk) 09:23, 13 September 2025 (UTC)

Upon rereading @Keith's post, I think maybe I see a bit of how we see these issues differently. I see all of the issues I mentioned through the lense of how they affect the general public. (Whether users, or non-users who are affected by AI or its uses.) Therefore, I see them all as "consumer" issues. Drakeula (talk) 10:21, 13 September 2025 (UTC)

Yeah, I think the primary point of difference we have is over that definition of 'consumer'. For me, I think if we don't keep that to just being the consumer-company relationship, it would result in pretty major scope creep for the wiki overall (you can imagine how many things could be brought into scope across a wide range of circumstances). There might be some flexibility in the definition of 'consumer' (e.g. I think you could potentially count small businesses as consumers in a context where they're buying a major peice of equipment from a supplier, and the supplier is messing them about), but I think for the wiki it always has to be 'someone who is purchasing products or services (paying with either their money or their data in the case of social media etc.), and is being subjected to unfair terms when doing so.

I'd absolutely appreciate further input on this, as I don't think it's a completely settled issue for the wiki, and the boundries of what is and isn't in-scope are not firmly drawn yet. Keith (talk) 13:27, 16 September 2025 (UTC)

I think @Drakeula is making some important points. I think it will take a lot of work, but this article can be restructured to be within scope. The core issue I see with it as of now is broadness and generalized issues. If this theme article could be structured around real incidents and very specific issues and such, it makes for a good article. I can work on it a bit to get it up to a higher quality standard, at least reworking the structure itself (and the text secondary) Beanie Bo (talk) 13:37, 16 September 2025 (UTC)

A hypothetical question on the scope: What about the relation between a citizen and a government, government agency? Would that be in scope or out of scope for the wiki? (I couldn't find anything that said one way or the other about this in the guidelines. If this is spelled out someplace already, maybe it could be made clearer in the inclusion guidelines.)

To me, governments (and govt. agencies) seem similar to corporations, but obviously there are also differences. (Would a church or a university count as a corporation?)

Presumably an issue between a private utility and customers would be in scope. What about a public utility?

Another example, Albania just appointed an AI as a "minister" to oversee procurement. If that AI were to do a rug-pull for example on a bunch of individuals, how would one figure out the scope?

I am inclined to think including governments in addition to corporations may make sense. But I does that open up another hairy can of worms? Drakeula (talk) 03:12, 17 September 2025 (UTC)

@Keith for the above. Drakeula (talk) 03:15, 17 September 2025 (UTC)

The more I look, the more I am getting confused about scope. Looking at the listing of Mr. Rossman's videos on this wiki Louis Rossmann - Video Directory (many of which seem to be prompting having an article made relating to them), I clearly don't see what makes something a company & consumer issue.

Specifically: Items about stripe, square and visa charging fees to small businesses (maybe I can see this, if small business is considered a gray area). Bunch of things about military and right to repair, (where is the consumer in that?). John Deere vs. farmers - what fraction of that is is individual consumers (as compared to small/midsized/large business).

Note, my intention is not to criticize here, I appreciate the work people are putting in on this. I am not saying any of the above should not be here, I am just trying to figure out why those are okay, but not some of the items above.

For example, what about DRM where Medical ventilators have to be blessed by factory authorized dealer? (Which was an issue at the beginning of covid, when suddenly a lot of ventilators were needed.) [This affected patients ("consumers"), but it is nominally between a business and a hospital (i.e., a business or governmental entity).]

If this needs discussion, it should probably be in some more visible talk page. Feel free to link to a better spot, or I will look for a better spot to ask when I get time. Drakeula (talk) 08:20, 17 September 2025 (UTC)

I agree completely that we need more clarity, and it's not something I have a fully formed idea of - it's conversations like these that will be essential to working out exactly where we draw the line. I'm starting to lean towards the ides of classing a 'consumer' as 'the buyer in a buyer-seller relationship, where an individual buyer, or the users of a purchased device, does/do not have sufficient leverage to affect the practices of the seller'.

--

The video directory is more of a 'here's a bunch of stuff, some of it will probably be relevant' situation, so it shouldn't be automatically expected that what's on there is in scope. Keith (talk) 10:09, 17 September 2025 (UTC)

For the question on Governments, maybe a better way to look at it is a consumer-seller relationship, rather than explicitly a consumer-'company' one. I think any rule that would allow an article about a private utility doing something, but not a public one, would be silly.

This pulls into the question, though, articles like Flock License Plate Readers. My 'gut feeling' is that this is an article that belongs on the wiki, however I acknowledge that the framework I've constructed would most likely exclude it (which probably means the framework needs adjusting, or clarifying). Keith (talk) 23:59, 19 September 2025 (UTC)

Control of information. Access to information is one of the central pillars of right to repair. Since AI radically changes what information people access, I am missing why it is out of scope.

AI summaries in search lead to: Loss of independent journalism.The Media's Pivot to AI Is Not Real and Not Going to Work Loss of review sights. Reviews provided by AI regurgitate manufacturer specs, give incorrect information about products, give questionable recommendations.Beware of the Google AI salesman and its cronies I haven't seen sources on this, but I see no reason why sites that post repair information/fora would be exempted from this widespread pattern.

While not as far as I know related to AI, Mr. Rossman recently noted that his repair business is no longer listed on Google. He said that is a big deal for the size of a repair business like his. That may make it harder for people to exercise the right to repair. Drakeula (talk) 09:23, 13 September 2025 (UTC)

Definitely consumer concerns and therefore I think worth mentioning. The cost of AI is largely hidden. Simply acknowledging that these are consumer concerns with AI, and pointing to good places to get started with the issues. In an overview article, if we only list "new" consumer concerns, it may make things needlessly difficult for the reader.

Data Centers. Electricity rates are projected to go up precipitously because of AI. Pollution and greenhouse gasses affect everybody. (MS, Google missing carbon reduction goals because of AI.) Tax subsidies are paid for by broad segments of the population.

In siting an AI data center, the local commissioners and the company typically make the secret deal, where the company gets public subsidy of electricity, taxes, water, etc., and the community gets a few jobs (at several million per job), plus whatever other arrangements. Even though consumers are excluded from the negotiations, that is still a consumer issue, since consumers pay the electricity rates, they breathe the air polluted by the on-site gas generators, they have their wells polluted by excessive water draw, they deal with noise pollution, the community has to deal with the electronic waste, etc. Drakeula (talk) 10:23, 13 September 2025 (UTC)

Labor concerns. People make choices about clothing brands based on how the people making the clothes are treated. Why should it be different with AI. Just because I am not the person getting [bleeped] by the corporation, doesn't mean it is only between the business and the employee. Sure, not everybody cares, but I think clippies might. Drakeula (talk) 09:46, 13 September 2025 (UTC)

Intellectual property Perhaps I could have picked a more evocative title? For the purposes of AI, every person is a "creative." This concerns use of your e-mails/facebook posts/tweets/photographs/security camera footage/footage from your roomba/footage from your cars cameras/etc. for training AI. Ownerships of the logs of where you go, what you buy (and where and when). Who you communicate with. Ownership of your medical scans. The output of your medical monitors (CGM, CPAP, pacemaker, smart watch, etc.), and their use in training/advertising/whatever. Publication of your AI queries, your chat logs. Use of AI responses. Drakeula (talk) 09:58, 13 September 2025 (UTC)

I suppose "big data" might be a useful adjunct term to use here. Drakeula (talk) 10:03, 13 September 2025 (UTC)

I think the way in which the terms of many services unreasonably opt you into AI training off your private data without properly, where it is framed as an anti-consumer contract term, is reasonable for inclusion. I'm less sure about the blatant grabbing of works visible in the public space, as this doesn't feel like as much of a consumer issue.

In a similar way, if there was a spate of phone repair companies using bullshit terms to refuse to return phones sent in for repair, that would fit the wiki, but if an organized crime gang was running a phone theft ring across a country, that would not fit the wiki. Both are stealing your phone, but the way in which it is being done is relevant as to whether it should be included in the Wiki. Keith (talk) 13:31, 16 September 2025 (UTC)

without properly informing the user or giving them a choice* Keith (talk) 14:32, 16 September 2025 (UTC)

On second thought, after re-reading this article in more detail, it seems like the only argument being made here is about unethical web scraping, which has no direct harm on consumers. Like you said, there are specific ways in which AI/LLM can be inherently anti-consumer, but that leans more toward specific incidents and company practice.

In summation, I do not think this article can be salvaged, and I think it should be deleted. Beanie Bo (talk) 14:23, 16 September 2025 (UTC)

I want to work on an article for the wiki about AI/LLM someplace where the whole business (talk page and article) isn't imminently going to disappear.

If by deleted you mean, you want to overwrite everything in this version of the article with something different, I have no particular opinion about that. (If bebold is a thing here, then bebold.)

If there had not been an article on AI, I would have started a draft on my personal page (if one can do that here). Since there is something here, there may be value in continuing it where multiple people can contribute.

If you want the article and talk page and all deleted, I would prefer not. As an alternative to deletion - Is there some way to mark this article as a draft/needing significant revision? (Most of my wiki experience is on wikipedia. I am still trying to pick up the policies/systems/etc. here.) Drakeula (talk) 02:29, 17 September 2025 (UTC)

Sorry, I should've been more descriptive. The way I see it, an article trying to outline why AI is bad or anti-consumer is like trying to say how social media is anti-consumer, or the internet is anti-consumer, etc. They're all simply too broad to write about. And, in reality, they're just tools. Social media can be anti-consumer in many ways, but it can be pro-consumer as well. Same with AI.

The only way I could see this article remaining is if it narrows the scope to common practices found all across AI/LLM's, which would only happen if incidents are compiled from ChatGPT, Claude, etc.. and even then, that's me being optimistic about it. I still can't see that working out in the long term. This wiki is based on advocacy, which means that the articles have to specifically detail events or practices that violate consumer rights in a tangible, direct way. Not indirectly. Not abstractly. Things that can be pointed to in a court of law.

Keith is the admin here, so I defer to him of course. But what I expressed above is just my personal take on it as a moderator and from what I've grasped from the other moderators, from the wiki policy, from Louis himself (in videos), etc. Beanie Bo (talk) 04:13, 17 September 2025 (UTC)

(also @Drakeula) I think there can be a place for an AI article, and there's definitely consumer-relevant stuff to talk about, and we certainly shouldn't do anything that would nuke this talk page (as I imagine this is a discussion which will come up again). For now, I'd say having this be a large page where we include all the broadly AI-related consumer protection issues makes sense, and it serves as a good place to discuss which sub-sections ought to be included. It might be that later on, the 'AI' page gets reduced to a very short page which mostly serves to act as a hub between various, more specific, pages on things like LLM platforms, upselling of AI integrations, whatever else - but it will need to be a big page before it can become a small page (if that makes sense).

@Beanie Bo I think any article like this would be classified as a 'theme' article, which can address broader trends without invoking specific examples at every stage (though citations and examples are, of course, always welcome).

Regardless, I think the overarching objective of this article shouldn't be "an article trying to outline why AI is bad or anti-consumer" (not saying that's what it is at the moment, just addressing Beanie's point), but rather "here is a collection of descriptions of the anti-consumer practices commonly associated with AI", with (for now) subsections that talk about different such practices, and eventually links to other articles that go into those areas in more detail. I think people are often going to want to click on an article called "AI", and that this would be the best way of making an "AI" article useful and informative to a reader without straying from the scope of the wiki. Does that sound sensible? Keith (talk) 07:27, 17 September 2025 (UTC)

@Keith Sounds sensible to me. Pretty much what I was thinking. Here is what might help a consumer understand as background to specific cases about "AI". Drakeula (talk) 08:28, 17 September 2025 (UTC)

Appeal posted re proposed deletion

I posted an appeal request regarding article deletion. Things I would like to see in this article are listed under scope. I think easier to edit existing article than start new one. If you have an opinion one-way or the other, please add on to the appeal discussion on the moderator page (wouldn't want multiple appeals). Thanks. Drakeula (talk) 18:41, 4 September 2025 (UTC)

Sources - potentially useful

General sources:

Pivot to AI. https://pivot-to-ai.com/ By UK journalist David Gerard. Daily news item about AI. He wrote about cryptocurrency. He noticed that many of the same people who had been promoting crypto took the same pitches and applied them to AI. So he similarly pivoted to AI.

Ed Zitron's Where's Your Ed At https://www.wheresyoured.at/ Journalist and PR professional, reports on financial aspects of AI.

Drakeula (talk) 18:15, 13 September 2025 (UTC)