AI startup Perplexity wants to disrupt the search sector. Forbes says he’s ripping them off


Artificial intelligence startup Perplexity AI has raised tens of millions of dollars from Jeff Bezos and other high-profile tech investors for its mission to rival Google in the information search space.

But its AI-powered search chatbot is already facing challenges as some media companies object to its business practices. It also competes with t I giant Google, and now Applewhich are increasingly merging similar AI capabilities into their core products.

Perplexity CEO Aravind Srinivas spent much of the last week defending the company after it published a summary report containing information and wording similar to a Forbes investigative article, but without citing the media or ask for permission. Forbes said he later discovered similar “counterfeit” stories from other publications.

The Associated Press separately discovered another feature of the Perplexity product making up false quotes from real people, including a former Martha’s Vineyard city official falsely quoted as saying he didn’t want the Massachusetts island to become a destination for marijuana.

“I never said that,” said Bill Rossi, a former board member of the island town of Chilmark.

Srinivas told the Associated Press that his company tries to build positive relationships with news publishers to ensure their news content “reaches more people.”

“We can definitely coexist and help each other,” he said.

When interviewed by Forbes, he said his product “has never ripped off anyone’s content.” Our engine doesn’t train on anyone else’s content,” in part because the company simply aggregates what other companies’ AI systems generate.

“We are actually more of an aggregator of information and we deliver it to people with the right attribution,” Srinivas said. But, he added, “Forbes specifically pointed out that they preferred more visible highlighting of the source. We immediately took this feedback into account and updated the changes the same day. And now the sources are more highlighted.”

Perplexity also revealed this week that it was seeking revenue-sharing partnerships that would pay news publishers a portion of Perplexity’s advertising revenue whenever a media outlet’s news content is referenced as a source.

Randall Lane, chief content officer at Forbes Media, called the dispute an “inflection point” in the AI ​​conversation.

“It’s a case study in where we’re going,” Lane told the AP. “If the people leading the charge don’t have a fundamental respect for the hard work of doing exclusive reporting and keeping people informed with value-added content, we have a big problem.”

Lane describes himself as an “AI bull” and believes the technology could help make many news organizations more efficient. Lane said the dispute between Perplexity and Forbes is important because it is a “metaphor for what can happen if the people who control AI don’t do it.” I don’t respect the people who do the work.

Perplexity presents itself as a search engine while “acting like a media company and publishing a story” that only Forbes had reported, Lane said.

“It was all very misleading. And what we didn’t hear was, ‘Oops, yeah, we blew it and we need to do better,'” he said. “Instead, it was simply a matter of publishing more content, small tweaks to the model, and treating journalism as if it were simply a product to be manufactured.”

Srinivas, a computer scientist and former AI researcher at OpenAI and Google, co-founded Perplexity in the summer of 2022, shortly before OpenAI’s Stable Diffusion and ChatGPT AI image generator began to attract public fascination for the possibilities of generative AI.

Inspired, in part, by his childhood love of Wikipedia, he described Perplexity to the AP as “like a marriage of Wikipedia and ChatGPT” that can answer a person’s questions instantly without the “huge mess” of conventional Google search results.

“You ask a question, you get an answer from clear sources, and there are about three or four suggested (follow-up) questions and that’s it,” he said of Perplexity. “This way, people’s minds can be freed from distractions and they can just focus on learning and getting deeper.”

The company sells a subscription for premium features and plans to launch an ad-based service as it expands its user base.

“We are not profitable as a company today, but we are also run more sustainably than model foundation companies, because we do not form our own model foundations”, which requires enormous amounts of computing power, he said.

Perplexity builds on existing large AI language models such as those built by OpenAI, Anthropic and Meta Platforms, Facebook’s parent company; then “post-train” them.

“We shape them to be very good summarizers,” he said.

It is not always clear where the summarized information comes from. A Perplexity feature called Writing – which allows a user to “generate text or chat without searching the web” – produces long, unsourced comments, often in the style of a news article. Testing of the feature by an AP reporter asking him to write about lack of marijuana on Martha’s Vineyard led him to produce a 465-word document that resembled a news article and included fabricated quotes from the former city official and another real person.

The AP does not repeat false quotes to avoid perpetuating misinformation. Srinivas said Perplexity’s writing feature is a “minor use case” intended to help write essays or correct grammar when primary source information is not needed. He said he was “more prone to hallucinations” – a common problem with AI large language models, as they are not tied to the web search capabilities of Perplexity’s main product.

“There is no doubt that generative AI is disrupting journalism, content creation and research,” said Sarah Kreps, director of the Tech Policy Institute at Cornell University.

She cited as an example Google’s new approach, such as Perplexity, which summarizes answers based on information extracted from web crawling. This too led to false information and forced Google will make adjustments to the product after its public release.

“But their entire advertising model is based on sending people to websites,” she said in an email. “Why will people go to websites if they can have a one-stop shop to find the answer in the AI ​​results? »

Srinivas told the AP that “a lot of people are getting referrals from Perplexity, and I’m happy that they’re getting referrals from a new player on the Internet.”

For now, many of these benefits could be aspirational. Perplexity’s global user base has grown rapidly this year, reaching more than 85 million web visits in May, but that barely compares to the billions of users of ChatGPT and other popular platforms from Microsoft and Google , according to data from Similarweb.

The debate demonstrates the “uncertain and difficult times” for online content creators in general and for journalism in particular, because aggregators only work if publications such as Forbes exist, said Stephen Lind, an associate professor at Marshall School of Business at the University of Southern California.

Using AI as a synthesis tool enables widespread dissemination of information until “you I’m running out of originals,” he said.

“There are entire companies or entire applications that are doing this as well, where they’re deploying new services without really thinking through the implications or the best practices or the safeguards, because they’re deploying applications for industries that they’re not in.” maybe not native,” he said.

Lind said it’s good that companies like Perplexity “take at least some course-correcting action when an industry or user backs up.” But some changes should have been incorporated from the start, he added.





Source link

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top