Reviews

About a month ago, the FTC announced a proposed rule that would ban using fake positive reviews and suppressing honest negative reviews.

Buried in it is one specific ban: “Businesses would be prohibited from creating or controlling a website that claims to provide independent opinions about a category of products or services that includes its own products or services.”

A lot has been written this year about how generative AI will impact marketing. More copy, more images, more assets. Faster.

And while it is possible this could be used for good marketing, it’s more likely it’ll end up being used for stuff like this: “Write me a review of [this product] from the perspective of a [target demographic].”

Sound unreasonable?

We don’t think it’s far off. While marketers hunt for the next arbitrage moment, running the content engine at full blast will likely become the “arbitrage” moment: put so much content into the world so quickly and so cheaply that you’re bound to capture attention.

This would be SEO to a new degree. But it’s also about something larger.

The goal of search to date has been to organize information and provide the most accurate results. But it hasn’t ever really been about the “best” results. Quality is subjective, hard to measure, etc.

Any filtering to address quality mostly happens on the search engine side, seems to be limited to anything that universally is considered “evil” or anything that so blatantly is exploitative of the algo that no human would ever bother reading it. Human users are also so lazy that further filtering of that information—to get better results—is near non-existent, even though Google has provided us with tools to do that.

For commerce, this has been a boon.

Search for “best ______” and the top results are all affiliate plays pushing an overlap of whoever has the best payout terms and the brands most likely to resonate with the publisher’s audience. You maybe can trust search to provide sites that will provide an accurate answer (you are, after all, getting results that someone is suggesting is a rundown of the best _____), but you can’t trust the sites to provide unbiased content or to not be participating in a “hidden” economy.

This is true for literally everything in commerce (B2B SaaS blogs, anyone?), not just consumer. Search is nearly, if not fully, trustless when it comes to commerce.

At some point, producing more untrustworthy stuff won’t really move the needle anymore.

At that point, ChatGPT’s influence ultimately may not be on creating more trustless content, faster—though it’s certainly going to do that. But what impact will that really have?

The content will be the same. It’s just an accelerant to a result.

So, what influence might ChatGPT (or elements of ChatGPT) have on search?

We think the real impact may be on building a model for trusted search.

Generative AI could, theoretically, change the trust paradigm of search (especially for commerce) by building a model that filters out sites with a likely vested interest. (Query the results, check the site and all links, run a search on the companies associated with the site to see what category they’re in, etc. Toss out anything that’s “commercial.”)

Maybe it’ll be able to get rid of fake reviews, too.