Missing links: Where are sources in Google’s AI search results?

Google’s introduction of Bard yesterday was huge news.

But perhaps the bigger news for SEOs was the preview of Google’s generative AI search results, which featured a grand total of zero links.

That’s right. Google – a search engine – no longer links to its sources.

You know, websites.

What do you call a search engine that copies mass amounts of content from others to create its own content, without providing any credit (links)?

I believe we still call that a scraper site.

A popular 2014 tweet about Google being a scraper site remains maybe even more relevant today:

As I’ve been digesting all the AI in search results news, the Harvey Dent quote from “The Dark Knight” popped into my mind: “You either die a hero, or you live long enough to see yourself become the villain.”

Now, this isn’t going to be another rant about Google being evil.

However, Google is playing a delicate game here.

Glenn Gabe called this an “act of war against publishers” on Twitter. And he’s not wrong.

For 24 years, Google has provided answers to queries in the form of a list of websites (or in carousels, or some other SERP format).

And over that time, user behavior has changed.

People want instant answers. We don’t always want to wade through 2,500-word articles when a question can be answered in 2-3 sentences or paragraphs. We’ve all complained about having to scroll past the epic backstories that precede nearly all recipe content.

Google has increasingly provided those answers in the form of search features outside of the organic results – knowledge panels, clocks, calculators, sports scores and many more.

The Google backlash got louder with the rise of so-called zero-click searches.

But now we may have zero-link search results. Really?

I remember when Google’s featured snippets made a lot of SEOs mad because they would “steal” traffic from websites. Others were mad before that when Google started showing the current time.

At the time, none of this concerned me. After all, you can’t copyright the time.

If your entire business model was to drive organic traffic from Google to your website just to tell people what time it was, you had a pretty flawed business model.

But here, Google may be entering a dangerous game.

Again, for two decades, websites have fed their content to Google, with the understanding they would get clicks (or at least visibility) out of the deal.

If Google just essentially summarizes (or rewrites, or whatever you want to call it) answers, what separates Google from being a scraper site now?

Or is it just straight up plagiarism?

Ryan Jones has an interesting Twitter thread, which basically says “no” to this idea. Of note:

• “None of them directly take text from any specific webpage – and the ideas they present in these examples aren’t things anybody can ‘own’ or ‘copyright.'”

But what happens if publishers and creators stop feeding content to the Google AI machine? What would happen to Google’s multi-billion dollar advertising empire, the majority of which has been built by monetizing search results filled with links to websites?

Today, Microsoft previewed its ChatGPT version of Bing search. It includes plenty of links and citations to the websites where it generated its answers. (In this interview, Microsoft CEO Satya Nadella said, “Everything was linkable, and that’ll be the goal — whether it’s inside a search, in the answer, or even in the chat session.”)

Neeva, which in January incorporated generative AI into its search results, has figured out how to link to its sources. So has You.com.

Google needs to do the same.

As Lily Ray put it on Twitter, “Every other search engine figured out how to cite its sources except the one with 90% market share, got it.”