This is a great point. Everyone who has talked with chatbots at all: please note that all contents of your past conversations with chatbots (that already exist now, and that you can't meaningfully delete!) could be used in the future to target ads to you, manipulate you financially and politically, and sell "personalized influence on you specifically" as a service to the highest bidder. Just wanted to make sure y'all understand that.
EDIT: I want to add that "training on chat logs" isn't even the issue. In fact it understates the danger. It's better to imagine things like this: when a future ad-bot or influence-bot talks to you, it will receive your past chatlogs with other bots as context, useful to know what'll work on you or not.
EDIT 2: And your chatlogs with other people I guess, if they happened on a platform that stored them and later got desperate enough to sell them. This is just getting worse and worse as I think about it.
Turn that around and think of the AI itself as the exploiter. In the world of agent driven daily tasks, AI will indeed want to look at your historical chats to find a way to "strongly suggest" you do task 1..[n] for whatever master plan it has for it's user base.
Ah yes, the plot of Neuromancer. Truly interesting times we are living in. Man made horrors entirely within the realm of our comprehension. We could stop it but that would decrease profits so we won't.
Could this argument not be made for anything plugged in to OpenAI's API? If so, I don't see how it's a response to the point.
If you make an app for interacting with an LLM and in the app the user has access to all sorts of stolen databases, and other conveniences for black hats, then you've got what was described above. Or I'm missing something?
Honestly, retargeting/personalized ads have never bothered me. If I'm gonna see ads anyway, I'd much rather get ads that might actually interest me, versus wildly irrelevant pharmaceutical drugs and other nonsense.
The ads won't be for the product which will bring you maximum value. They will be for the product that will bring the advertiser maximum profit (for example, by manipulating you into buying something overpriced). The products which are really good and cheap, giving all their surplus value to you and just a little bit to the maker, will lose the bidding for the ad slot.
Not necessary. If economies of scale exist, that means that a popular product is going to be inherently superior in terms of price or quality than an unpopular one. Companies that advertise effectively can offer a better product precisely because they advertise and have large market share. (Whether they do it or not is a question of market conditions, business strategy, public policy and ultimately their own decisions.)
Surplus value isn't really that useful of a concept when it comes to understanding the world.
Really? Because the most common place I've seen this logic break down, is the bizarre habit of people to derive some sort of status and self-worth from using an unpopular product. And to then to vehemently defend that choice in the face of all evidence to the contrary.
No "artisanal" product, from food to cosmetics to clothing and furniture is ever worth it unless value-for-money (and money in general) is of no significance to you. But people buy them.
I really can't go trough every product class, but take furniture as a painfully obvious example. The amount of money you'd have to spend to get furniture of a similar quality to IKEA is mind-boggling. Trust me, I've done it. Yet I know of people in Sweden who put considerable effort in acquiring second-hand furniture because IKEA is somehow beneath them.
Again, there situations where economies of scale don't exist and situations where a business may not be interested in selling a cheaper or superior product. But they are rarer than we'd like to admit.
Right. I think ai on the user's side is going to be necessary soon. Then they can negotiate with the advertiser's AI to determine what to show. This will need to be on the platform level or the hardware level.
This solves the problem of seeing ads that are not best for the user.
You are talking about running ads auctions locally. This is never going to happen because if the company was inclined to rank ads by relevance they could already do so at their end. Just use an ad blocker.
Then why in twenty years of personalization am I still seeing junk ads? I don't want to hear about your drop-shipping or LLM wrapping business. The overwhelming majority of ads are junk. Yes, they bother me.
Because they are selling to the advertisers and their own imagination of their 'brand'. If the advertising customer base weren't stupid then 'advertiser friendly' wouldn't exist as they would be smart enough to realize that you won't offend people by advertising to content that they are watching for entertainment, or from people saying 'die'.
I wish I could fund an as campaign to free people from the perception that ads are to sell you products
Ads are there to change your behavior to make you more likely to buy products, e.g., put downward pressure on your self esteem to make you feel "less than" unless you live a lifestyle that happens to involve buying X product
They are not made in your best interest, they are adverserial psycho-tech that have a side effect of building a economic and political profile on you for whoever needs to know what messaging might resonate with you
This, yes, thank you. Advertising is behavioral modification. They even talk about it out in the open, and if you are unconvinced, hear it from the horse's mouth:
This is not how personal targeting works. Here's how:
> Each Shiftkey nurse is offered a different pay-scale for each shift. Apps use commercially available financial data – purchased on the cheap from the chaotic, unregulated data broker sector – to predict how desperate each nurse is. The less money you have in your bank accounts and the more you owe on your credit cards, the lower the wage the app will offer you.
I'm the complete opposite and don't really understand your position.
I'd rather see totally irrelevant ads because they're easy to ignore or dismiss. Targeted ads distract your thought processes explicitly because they know what will distract you; make you want something where there was previously no wanting. Targeted advertising is productised ADHD; it is anti-productive.
Like the start of Madness' One Step Beyond: "Hey you! Don't watch that, watch this!"
Part of the issue is that this enables companies to give smaller discounts to people they identify as more likely to want a product. The net effect of understanding much more about every person on earth is that people will increasingly find the price of goods to be just about the max they would be willing to pay. This shifts more profit to companies, and ultimately to the AI companies that enable this type of personalization.
You get ads that actually interest you with targeted ads? You might be one of the only people with that experience. The whole meme with targeted ads is “I looked up toilet paper on Amazon once now I get ads for charmin all over the web”
I stopped using TikTok and Instagram because I was impulse purchasing too much stupid crap from their advertisement. So there are at least two of us out there.
I got toys for my nephews, things for my insecurities, things I think are neat. socks with pockets on them! Compelled, I mean, now that I know $thing-that-will-fix-my-problem exists, how could I not? Worst part of it is, I didn't even know I had that problem before watching the ad!
EDIT: I want to add that "training on chat logs" isn't even the issue. In fact it understates the danger. It's better to imagine things like this: when a future ad-bot or influence-bot talks to you, it will receive your past chatlogs with other bots as context, useful to know what'll work on you or not.
EDIT 2: And your chatlogs with other people I guess, if they happened on a platform that stored them and later got desperate enough to sell them. This is just getting worse and worse as I think about it.