The new AI-based search engine has a gimmick: humans answering the questions

Furthermore, he claims that Perl is much less likely to deliver misleading information than many other AI search engines — which he believes are more likely to deal with a “tidal wave” of lawsuits based on the bad answers it provides. “These other players are building amazing technologies. Call them Ferrari or Lamborghini,” says Kurtzig. “We are building a Volvo – safety first.”

This idea of ​​Perl’s superiority, of course, made me more eager to try it. Kurtzig seemed certain that Pearl would still enjoy Section 230 protection. Amnesty International asked if she agreed.

It would likely qualify as an “interactive computer service” under Section 230, Pearl said, meaning it would be protected from being treated as a publisher, just as Kurtzig predicted. But, the AI ​​continued, “Perl’s position is unique because it creates content using artificial intelligence.” He didn’t have a specific answer for me after all.

When I asked to speak to a lawyer directly, he redirected me to JustAnswer, where he asked me to provide the answer I wanted to verify. I said I needed to go back and copy the answer, as it was several paragraphs long, but when I returned to the Pearl website, the conversation had ended and I was reassigned to a new chat.

When I tried again, this time opening Pearl on the desktop, I received a similar uncertain answer. I decided to conduct an examination of the human facts; A few minutes later, I got a TrustScore™ — 3!

Pearl recommended that I seek out an actual expert’s opinion, and sent me to her signup page. I got a login so I didn’t have to pay while testing the tool. She then connected me with a Legal Eagle expert.

Unfortunately, the lawyer’s answers were not clearer than the artificial intelligence’s answers. He noted that there is an ongoing legal debate about how Section 230 applies to AI search engines and other AI tools, but when I asked him to provide specific arguments, he gave a strange answer noting that “most of them use shell companies or associations to file.”

When I asked for an example of one of these shell companies — I was confused about how this related to a general discussion of Section 230 — the Legal Eagle asked me if I wanted him to put together the package. I said yes more confused. I got a pop-up indicating that my expert wanted to pay me an additional $165 to research the information.

I declined, frustrated.

Then I asked Pearl about the history of WIRED. The AI ​​response was serviceable, although it’s the same stuff you’ll find on Wikipedia. When I asked about TrustScore™, I again encountered the number 3, indicating that it was not a very good answer. You have selected the option to connect with another human expert. This time, perhaps because it was a media question and not a straightforward legal or medical topic, it took a while for the expert to appear — more than 20 minutes. When he did, the expert (it has never been confirmed what gave him media goodwill, although his profile indicates he has been with JustAnswer since 2010) gave me an answer remarkably similar to AI. Since I was taking a free test, it didn’t matter, but I would have been upset if I had already paid the subscription fee just to get the same mediocre answer from both a human and an AI.

On my last attempt to use the service, I asked a direct question: How to refinish kitchen floors. This time, things went more smoothly. The AI ​​returned a suitable answer, resembling a version of a very basic YouTube tutorial. When I asked the human expert to assign a TrustScore™, they gave it a 5. It certainly seemed accurate enough. But — as someone who really wants to refinish the old pine boards in my kitchen yourself — I think when I’m actually looking for guidance, I’ll rely on other online communities for human voices, ones that don’t charge $28 a month: YouTube and Reddit.

If you end up testing Pearl, or any new AI-powered search products, and had a memorable experience, please let me know how it went in the comments below the article. You can also contact me via email at kate_knibbs@wired.com. Thanks for reading, and stay warm!