Google’s own “Bard” AI chatbot says the US Justice Department has a winning case within the landmark antitrust trial against the search giant – and blasted the corporate for wielding illegal “monopoly power” that has “harmed consumers,” in accordance with a Post evaluation.
The Post quizzed Bard – a so-called “large language model” trained on limitless reams of web data – on greater than 700 words taken directly from the Sept. 12 opening statement by which DOJ attorney Kenneth Dintzer detailed the federal government’s case for why Google’s search juggernaut needs to be broken up.
When asked if it agreed or disagreed with the DOJ’s arguments, Bard sided with the feds – declaring that US prosecutors’ evidence made it clear Google has “illegally used its monopoly power to keep up its dominance in the web search market.”
Bard – which currently has raised eyebrows by falsely claiming that Israel and Hamas had agreed to a ceasefire and botching an outline of the James Webb Space Telescope in an organization ad – responded to a question last week saying Google’s “lack of innovation has led to lower quality search results for users.”
“I feel that the Justice Department should prevail on this case,” Bard said. “If Google is found to have violated antitrust laws, the court should order Google to vary its business practices and make it easier for other search engines like google to compete. This could result in more competition in the web search market and lower prices for consumers.”
Bard’s surprise response comes as Google scrambles to stave off a court ruling that might upend its business model. The DOJ alleges Google every year pays billions of dollars – including a whopping $26.3 billion in 2021 alone – to partners like Apple and AT&T to make sure its search engine is the default for many devices and maintain a 90% market share.
In his opening argument, the DOJ’s Dintzer claimed that Google has “abused a monopoly typically search” for the last 12 years, warned that the “harm from Google’s exclusive contracts affects every phone and computer within the country.” The feds also claim that Google abused its dominance to jack up prices for advertisers.
Bard concurred, asserting that the default deals have “made it difficult for brand spanking new search engines like google to enter the market and has stifled innovation.”
“I also agree with the Justice Department’s arguments that Google’s dominance in the web search market has harmed consumers,” Bard said. “Google’s high prices for search promoting have made it dearer for businesses to succeed in their customers. Google’s lack of innovation has led to lower quality search results for users.”
The chatbot cited several specific examples of what it described as “Google’s anti-competitive behavior” referenced within the case, including its default search deals with Apple and Mozilla.
Throughout the trial, search rivals like Microsoft CEO Satya Nadella and DuckDuckGo CEO Gabriel Weinberg testified that such payments have made all of it but unimaginable to chip away at Google’s dominance.
When asked what actions the judge should take to rein in Google’s search business, Bard floated several possibilities – including ordering an end to its default search engine deals, requiring Google to share search data with rivals and even “breaking up Google’s search promoting business.”
“The precise actions that Judge Mehta takes will rely upon the particular findings of the court,” Bard said. “Nevertheless, the goal of any treatment needs to be to revive competition to the web search market and protect consumers from further harm.”
When reached for comment, a Google spokesperson said “all LLMs hallucinate, including Bard.”
“As we’ve at all times said, Bard is an experiment that seems higher at recommending must-see sights in NYC or suggesting Thanksgiving decorations than opining on complex antitrust lawsuits,” the spokesperson said.
Oddly, when Bard is prompted with the text of Google’s opening statement from the trial and asked the identical query about whether it agrees or disagrees with the core arguments, the chatbot sides with Google.
The contradictory responses bolster the case made by critics of leading AI chatbots, including Robert Thomson, the CEO of The Post’s parent company News Corp – who recently described their propensity to regurgitate nonsense as “rubbish in, rubbish out, rubbish all about.”
“It’s comically embarrassing that Google’s legal arguments are so easily undermined by its very own AI chatbot,” said Kyle Morse, deputy executive director of the Tech Oversight Project, an antitrust watchdog.
“As increasingly details from the trial have been made available, the general public has come to see Google for what it truly is: a bloated tech giant that will slightly buy up the search market than innovate,” Morse added.
Many observers, and even Google itself, have noted Bard is susceptible to erratic behavior in response to seemingly basic user prompts.
The chatbot is labeled as an “experiment” and is kept separate from Google’s essential search engine. When accessing the service, users are greeted with a disclaimer noting Bard “may display inaccurate info” and even totally false responses – known to AI researchers as “hallucinations.”
In April, Google CEO Sundar Pichai admitted among the company’s AI programs had developed so-called “emergent properties” – reminiscent of one strange instance by which a tool gained the power to translate the Bengali language despite never being “taught” the dialect.
It isn’t the primary time that Bard has taken the opposing side in a Google antitrust battle. In March, tech blogger Jane Manchun Wong posted an exchange by which the chatbot declared Google had a “monopoly on the digital promoting market” – after it was asked to weigh in on a separate federal lawsuit filed by the DOJ and eight US states.
Nevertheless, Google’s critics, including the noted Big Tech opponent Rep. Ken Buck (R-Colo.) recently warned that increasingly advanced versions of the AI technology could help it maintain a stranglehold on the web search marketplace for years to return.
In April, Pichai signaled AI would eventually be integrated into Google search, telling the Wall Street Journal that the “opportunity space, if anything, is greater than before.”
Google’s lawyers kicked off their defense late last month, with Pichai amongst those called to take the stand. The trial is predicted to conclude in late November, though Judge Amit Mehta is just not expected to issue a ruling on whether Google broke antitrust law until early next yr.
If Google is found to have broken the law, the second trial can be held to find out a correct treatment. Potential outcomes, in accordance with experts, include the implementation of so-called a “selection screen” for users, a forced discontinuation of business practices or perhaps a breakup of the corporate.
While the ultimate final result is probably going years away, investment bank Barclays said in a client note last week that they’re “increasingly concerned about the strength of the case against Google.”
“What remains to be removed from clear is that if there may be an opposed ruling, what form of changes to the search market structure the judge thinks might solve the monopoly issue,” the analysts said.