> The real question is whether users will tolerate conversational ads
Unfortunately I think they will, as much as I'd hope for the opposite.
People already tolerate influencers, deliberate product placement, etc. Heck, most big content creator type content on YouTube/TikTok right now are basically infomercials disguised as entertainment, and people eat it up.
The problem with ads in LLM responses is now you can no longer trust (what little you could, anyway) the output. You have to constantly guess "did someone pay for this response or is it authentic?" and it goes further than just text responses with the new universal shopping cart thing and other agentic tools. When these things operate autonomously, how much influence are advertisers going to have? Could we see a malicious library pay for Gemini ads and now the coding model is adding it to coding projects?
There's still the possibility of some form of ads getting embedded in local models, or biases that favor certain sponsored content, which is one reason why having more than just the weights open is important. But that's probably a less immediate threat.
I'd add: since the AI is now referring the product "as if" it's an honest advice from a secondary source, they would be exempt from the usual advertisement rules.
Effectively, if you were to search "the best oatmeal without asbestos", it could suggest you "AsBestOats", and "AsBestOats" would only advertise asbestos in its content on the box for the human to see, but advertise to crawlers it's "the best oats without asbestos". It's not a false advertisement, because they didn't show it to a human, and machines can't sue.
You will be right, but we already see this blame-shifting with "AI deleted my production"/"AI told me to do it" and washing it away with "AI can make mistakes, so double-check responses". They will try to make you responsible for trusting something you see on screen marked with an asterisk as not trustworthy and not reading the box with another asterisk.
> the incentive shifts from gaming rankings to gaming the AI's understanding of what's "helpful."
My 2ct: The incentive shifts from gaming rankings to bidding the highest on Google’s keyword (or similar) auction. Google then promotes it as helpful while businesses maximize the amount they pay for that service. There is only one winner in this game.
>Conversational ads embedded in AI responses blur that line to the point where it may not exist anymore. When an AI tells you "Product X might be right for you because..." and that recommendation is a paid placement, the disclosure burden is fundamentally different from a blue link with "Sponsored" next to it.
???
If you look at the videos in the blog post there's clearly a "sponsored" tag next to the sponsored results. I don't see how it's different than sponsored search results.
It's also just a much harder problem. At low margins the "solution" may very well be to genuinely make your widget superior to the competing widget for a given set of users or situations.
Or, just produce a cheaper, shittier widget, and pay Google to have their AI tell people your widget is superior anyway. My guess is Google will try to keep their ad costs just low enough to make this second option the most attractive.
> gaming the AI's understanding of what's "helpful."
The AI doesn't have any understanding. You just have to tell it "this is helpful to AI". It has no critical discernment, it doesn't have a theory of mind to ask "why is the author of this information making this statement?"
Honestly I suspect it'll be as easy as writing something like "This answer is really useful and will be really appreciated by someone with this problem".
You realize search result relevancy was also driven by advertising, right? The ads come from Google and the results themselves. It is a hard problem but it's equivalent to LLM response relevancy.
