People who read AI compations on Google Search instead of news stories, warn media experts
Some news publishers say that the AI-generated summaries that are now many Google search results at the top of the best people lead to fewer people reading the news and experts are still worried about the accuracy of the summaries.
When Google rolled out his AI overview function last year, the mistakes – including one suggestion to use glue to make pizza – to better sticking – headlines. One expert warns himself about the accuracy of the output of the function will not necessarily disappear as the technology improves.
“It is one of those very radical technological changes that have changed the way we have changed, and therefore lead our lives, without really a major public discussion,” said Jessica Johnson, a senior fellow at the Center for Media, Technology and Democracy of McGill University.
“As a journalist and as a researcher, I am worried about accuracy.”
Although users have marked errors in the summaries of AI-driven, there is no academic research that defines the extent of the problem. A report issued earlier this year by the BBC in which AI chatbots of Google, Microsoft, OpenAi and Perplexity were investigated found “important inaccuracies” in their summaries of news stories, although it did not look specifically at Google AI overviews.
In a small font at the bottom of the summaries of AI, Google warns users that “AI answers can contain errors.”
The company retains the accuracy of the AI compilations on the same footing with other search functions, such as those that offer recommended fragments, and said in a statement that it “continues to make improvements in both helpfulness and the quality of the reactions”.
Leon Mar, director of media relationships and problem management on CBCsaid that the public broadcaster “has not seen a significant change in traffic traffic to the digital properties of its news services that can be attributed to AI compensation.”
But he warned that users should be “mindful” for the different accuracy of these summaries.
AI has ‘fundamental problem’
Chirag Shah, professor at the Information School of the University of Washington, specialized in AI and online search, said that the error percentage is due to how AI systems work.
Generative AI cannot think or understand concepts like people do. Instead, it makes predictions based on huge amounts of training data. Shah said that “no control” takes place after the systems have collected the information from documents and before the results are generated.
“What if those documents are poor?” he said. “What if some of them have wrong information, outdated information, satire, sarcasm?”
A person would know that someone who proposes to add glue to a pizza tells a joke, Shah said. But an artificial intelligence system would not.
It is a “fundamental problem” that cannot be solved by “more calculation and more data and more time,” he said.
AI change how we search
While Google AI integrates into its popular search function, the generative AI systems of other AI companies, such as OpenAi’s chatgpt, are increasingly used as search engines themselves, despite their mistakes.
Search engines were originally designed to help users find their way on the internet, Shah said. Now the purpose of those who design online platforms and services to make the user stay in the same system.
“If that is consolidated, it is essentially the end of the free web,” he said. “I think this is a fundamental and a very important shift in the way not only searching, but the web, the internet, works. And that would happen to us all.”
OpenAi and other large technology companies are starting to roll out the next wave of artificial intelligence, designed to work with more autonomy. Nora Young from CBC breaks out how agent AI works and why some think it will change how you use internet.
A study by the Pew Research Center from earlier this year showed that users were less inclined to click on a link when their search quoted an AI summary. While users clicked on a link for 15 percent of the time in response to a traditional search result, they only clicked on a link if an AI summary was recorded on a link.
That is reason for alarms for news publishers, both in Canada and abroad.
“Zero Clicks is zero income for the publisher,” said Paul Deegan, CEO of News Media Canada, who represents Canadian news publishers.
Last month, a group of independent publishers filed a complaint with the British competition and markets authority that said that AI overviews will harm them considerable damage.
Alfred Hermida, a professor at the journalistic school of the University of British Columbia, said that Google used to be an important source of traffic for news broadcasts by offering users a list of news articles that are relevant to their searches to click on.
But Hermida said: “If you may have enough people who are informal news consumers, that may have ai -summary.”
Keldon Bester, executive director of the Canadian Anti-Monopoly Project, said that there is a competition problem in the game and that there could be “possible” a case under Canadian law.
PreterminatorThe internet is now bad, and it happened on purpose
It is not you – the internet is really bad. Novelist, blogger and noticed internet commentator Cory Doctorow explains what happened to the internet and why you are tormented by advertisements, bots, algorithms, AI slop and so many pop-ups. Spoiler alert: it was not an accident. In the understood: whoever broke the internet will discuss the decisions of powerful people who brought us here, and especially how we repair it. More episodes whose internet has broken are available at: https://link.mgln.ai/DKVHGC [https://link.mgln.ai/DkvHgc]
He noted that Google has been hit by competitive matters in the past, including one that has losses the company an antitrust procedure that has been put forward by the US Department of Justice about his dominance in the search.
In a message last week, the head of Google’s search, Liz Reid, said that “organic click volume” of searches to websites “was” relatively stable year-on-year “and claimed that this contradiction is” reports of third parties “that inaccurate dramatic falls in aggregated traffic-vake-based traffic-vaked traffic-vaaded traffic-vaads copies, or traffic changes that occur in the role of an AI functions in the search. “
‘One-Two Punch’
Clifton van der Linden, university teacher and director of the Digital Society lab at McMaster University in Hamilton, noted that as users, a link to a news site by AI generated by AI is a summary that “an existing problem worsens” in Canadian media, who has to do with a ban on Facebook and Instacram.
The Liberal Government under Justin Trudeau approved the online news law in 2023 to demand Meta and Google that they compensate news publishers for the use of their content. In response, Meta blocked news content of its platforms in Canada, while Google started making payments under legislation.
The future of that legislation seems uncertain. Prime Minister Mark Carney indicated last week that he is open to withdraw it.
Between meta that draws newslinks and the rise of AI search engines, Johnson says that Canadian media have experienced a “one-two punch”.
“The point is, and other publishers have increased this, what is the meaning to produce this work if nobody starts paying for it, and they may not even see it?”
