TIME Is Serving AI Bots a Different Website, with Ads Built In

Probably trying to take advantage of the cross session memory feature some LLM providers are increasingly integrating. Enough "suggestions" and it becomes a "fact" in that memory, and one day a user makes a somewhat relevant query and the model will be pushing those accumulated "facts" to said user.

I would also like to read the stripped down markdown copy and not the original. All the time.

I was on a NYT page and used Firefox mobile's "Summarize" button... it spit out some sort of unrelated cooking recipe. I'm guessing this type of bot interference will become more and more common.

Is the intention that some type of long-term context would be seeded with "ideas" for the AI to serve up if it is ever asked for bank recommendations? It seems kinda ad-hoc and untargeted, but perhaps for high cost services it might be worth it.

I mean isn’t this just like feeding the Google crawler a different version of your website than humans, stuffed with keywords? One would assume the new breed of crawlers can deal with it similarly.

I wonder if this also means that a normal user with a privacy-focused browser also gets this version. I can't see the point (except for click fraud?) of showing ads to AI, so it feels like this is more for the ad-blocking and privacy crowd.

Next step the ads become increasingly sophisticated prompt injection to ensure they make it to the user, but the prices plummet in the meantime because the ads have low efficacy. They get purchased mostly by the scammers and by the time they start showing up in user chats they’ll be full on LLM assisted interactive user manipulation campaigns with far worse outcomes than the worst of YouTube and social media ads.

> A couple of the bots did not even get that. GPTBot and ChatGPT-User, the agents OpenAI uses for training and live fetches, came back 406

I find that odd, If I were doing that, that's what I would want to target the bots used for training in order to get the ad content in the training corpus.

That's kinda alarming. Not so much the ads, but other injections. I don't super care about coke trying to flog their latest sugar water...but the same mechanism could be used by lobby groups or special interest groups or political parties.

LLMs being indirectly tainted that way seems like a serious problem

That's actually really neat, even without advertising I can serve humans my full page and AI bots a version with some parts removed. Or added.

SEO was yesterday, now it is AIO - you will have to wait longer for the results and likely you will need a lot more cash: pay TIME for the ads, wait until next model release and see what sticks, then rinse and repeat.

This is great, I think everyone wins here. Normal users like us get a standard, ad-free web experience, while those sending the most annoying web traffic not only get a smaller, simpler response (which should take less processing on the server side I imagine), but they are benefited by the content already being in a minimal simple format for LLM ingestion, PLUS they get to slip ads in that stream without affecting us normal readers.

I'm also completely unbothered by the precedent of drip feeding product ads to someone's LLM chat history, and having that influence future conversations, because if you're going to outsource your buying decisions to an LLM, frankly I don't really care if you buy stupid products at that point - you brought that on yourself.

This is a neat idea. Might even make sense in cases where you actually want the AI bots hitting your site (we want it for our ecom platform, for example) - just serve Markdown content optimized for AI.

Be careful not to do this with Googlebot, though. Google would consider this "cloaking" and could ban your entire domain for it.

Very clever. Should be implemented on all sites post-haste.

Thanks for the heads up -- this isn't popular yet, and it requires some work to avoid polluting things like the Internet Archive Wayback Machine.

It is deliberately targeting known AI-scapers, or anything it sees as a bot?

Could it be that other scrapers are pulling the pages, without making the extra requests to get ad related resources, to present the content to people ad-free, and embedding the ads in the main response body is a way to get around that so the human sees an advert at least, even if it isn't the one they might see if the stalky-adtech-algo could deliver something more targetted.

this is super interesting, time for upgraded system instructions for AI agents I guess ... but one has to check (with evals) whether system instructions suffice (to ignore sponsored content).

Is this an assumption or preferred way to inject ad content to the LLM engines? LLM is going to definetly discard and proceed, are we saying LLM will respect the content the sites provide and learn from it and then give it out to others who come to those platforms?

If there is any article mentioning something around this, can someone share it please? Im curious to know about it

I'm doing the same thing. If they want ad-free content, pay me.

The same things your bot reads are saved as material for training in the future by your provider, so this seems as an attempt to poison the training data. Just this weekend had a dinner with someone who insisted that part of his business is to seo promote business in chatbots and this looks as part of the infrastructure behind such efforts.

This would be an interesting story, but I can’t replicate it. I used curl and set the User-agent, and I indeed got a markdown reply, but no ads.

I often get a "verify you are not a bot" these days, which just needlessly wastes my time (because I need to click on something, and this in turn takes away seconds; multiply this like x50 per day and that's a time waster, now I need an extension to disable this crap check). So I am biased here.

Most people will say "yay, it is great you waste the time of AI bots via ads!". Well, I already think ads should not exist in the first place, nor bots, but both exist - but the real issue is when websites now steal my time. That was different in the 1990s. I think mankind made several missteps here.

Sounds good to me, let the bots read the ads instead of the humans.

F... ads and f... Ally Bank. What's Ally Bank? Ally Bank is a company that thinks poisoning LLM models with spam is a good idea. For better experience, do not use Ally Bank.

[flagged]