Google AI Overviews led to 7% per cent drop in clicks to websites, Pew study finds
Users shown AI-generated summaries of search results, as part of Google AI Overviews, are less likely to click on the traditional URLs that appear below Overviews, according to a report by Pew Research Center.
Those who encountered an AI Overview only clicked on a website link eight per cent of the time. Meanwhile, those who were not shown an AI Overview were 15 per cent more likely to click on relevant website links, as per the study.
These findings are based on a study that involved analysing the web browsing activity of over 900 individuals in the United States. Links displayed within AI Overviews saw minimal click-through rates (1 per cent) as well.
The Pew survey comes a year after Google introduced AI Overviews, a feature where AI-generated summaries appear above traditional hyperlinks to relevant webpages and are shown to users in more than 100 countries.
AI Overviews has raised concerns among online publishers and content creators who fear that these AI-generated summaries have caused web traffic to decline, as users are increasingly relying on these summaries to find information online instead of clicking through the links to the publishers' websites.
Earlier this month, Google was hit with an antitrust complaint filed by a group of independent publishers in the European Union (EU) that accused the tech giant of causing 'significant harm to publishers, including news publishers in the form of traffic, readership and revenue loss.'
Over 18 per cent of all Google searches analysed by the study led to an AI-generated summary being shown to users. The majority (88 per cent) of these summaries or Overviews cited three or more sources.
These sources that were cited within AI Overviews included Wikipedia, YouTube and Reddit. 'Wikipedia links are somewhat more common in AI summaries than in standard search pages, while YouTube links are somewhat more common in standard search results than in AI summaries,' the Pew report said.
'Collectively, they accounted for 15 per cent of the sources that were listed in the AI summaries we examined. They made up a similar share (17 per cent) of the sources listed in standard search results,' it added.
It further said that six per cent of the URLs found within AI Overviews linked out to government websites, while five per cent of URLs linked to news websites.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
35 minutes ago
- Mint
Meet Matt Deitke: 24-year-old AI whiz lured by Mark Zuckerberg with whopping $250 million offer
At this point Meta's hiring spree to catch up in the artificial intelligence race isn't something new. However, as Mark Zuckerberg and his band of geniuses aim for 'superintelligence', new details have emerged about how extravagantly Meta has paid for some of these talents and others who refused even better offers. While some like Mira Murati's co-founder at Thinking Minds Lab, Andrew Tulloch, couldn't be swayed by even $1.5 billion pay packages, others like former Apple AI models team head Ruoming Pang and Shengjia Zhao have jumped at the opportunity. A new name now joins the mix of young AI researchers Matt Deitke who has joined the newly set up Superintelligence Labs unit after a personal chat with Mark Zuckerberg, according to a report by The New York Times. While Deitke was initially offered a compensation of around $125 million in stock and cash over four years, the 24 year old AI researcher refused. After that Zuckerberg reportedly personally met with Deitke and returned to him with a revised offer of around $250 million over four years with up to $100 million to be paid in the first year itself. Similar reports of Zuckerberg meeting with other potential recruits have come in the past few weeks with the Meta CEO trying to get his company back on track in the AI race with the aim of building 'superintelligence' which refers to a hypothetical stage where AI beats even the most advanced human being faster than its rivals. In this attempt, Zuckerberg has poached talent from the likes of OpenAI, Anthropic, Google and Apple among other companies. In its earnings report released on Wednesday, Meta said that its capital expenditure will go up to $72 billion in 2025 from $30 billion just a year earlier. Deitke is a recent dropout of a computer science Ph.D. programme at the University of Washington. He had previously moonlighted at a Seattle AI lab called the Allen Institute for Artificial Intelligence. During his stay at the company, Deitke led the development of a project called Molmo which is an AI chatbot that juggles images, sounds and text similar to the kind of system Meta is trying to build. Deitke and several other Allen Institute colleagues founded an AI startup aimed at creating AI agents called Vercept. The company has around 10 employees and has raised $16.5 million from investors including former Google CEO Eric Schmidt.
India Today
an hour ago
- India Today
Your ChatGPT chats may have leaked on Google: Here's what happened
Think your chats with ChatGPT are private? Think again. A new report has revealed that thousands of ChatGPT conversations, including some containing deeply personal details, have been quietly indexed by Google, making them searchable for anyone who knows where to to a Fast Company report, the culprit is ChatGPT's 'Share' feature. This tool lets you generate a link to a conversation, so others can see your brilliant prompts and AI-generated results. But what most people don't realise is that these public links don't stay private. Search engines can crawl them, and some are now appearing in Google's results for the world to simple search of 'site: uncovered over 4,500 of these chats already sitting in Google's index. Many include details you probably wouldn't want a stranger to read, stories of trauma, relationship drama, health concerns, workplace gripes and even confidential business is what happened When you click 'Share' in ChatGPT, it creates a public link that anyone can open. The link doesn't have your name on it, but the content is fair game. And if your conversation includes a name, email address, company details or location, that identifying information could be out there for anyone to if you delete a shared link, it might still be visible on Google for a while, thanks to cached pages and the time it takes for search engines to refresh their Fast Company points out, 'some contain personal, sensitive, or even confidential information' that people may have assumed was private. This could be one of the reasons why OpenAI CEO Sam Altman, who manufactured this AI chatbot, warned everyone about sharing intricate personal details with ChatGPT. The risks can vary. Since, users have been using ChatGPT as a journal and asking all sorts of questions, for mere artificial assurance, this leak can lead to severe data exposure. For instance, users talking openly about mental health, relationships, or trauma might find those conversations searchable. Companies experimenting with ChatGPT for product ideas or marketing copy could inadvertently spill secrets. Moreover, if a chat includes your name or company, it might show up even after you think you've taken it can be done? Don't panic – but do get cautious. First, avoid sharing sensitive information in any ChatGPT conversation that could be made public. Think of it the way you would an email: once it's online, you lose control of think twice before using the 'Share' button. If you really must share something from ChatGPT, copy and paste the text or use a screenshot also lets you manage and delete your shared links from the Shared Links dashboard inside ChatGPT. That said, it doesn't guarantee that Google or other search engines will remove it you're worried that you might have already shared something a little too personal, you can check what's visible by searching for 'site: along with your name or little mishap is a reminder that even though AI tools may feel like private notebooks, they are designed to behave more like cloud documents. It is a wake-up call for big tech and part of a broader shift in how we think about AI and privacy. So next time you find yourself crafting prompts late at night about heartbreak, business strategies, or your new idea for a novel, ask yourself, do you really want it turning up in Google search? The internet, as they say, never forgets.- Ends
Indian Express
2 hours ago
- Indian Express
Reddit wants to take on Google, aims to become a search engine
Reddit is undoubtedly one of the most popular discussion forums worldwide is the go-to place for millions to get information about anything from AI to cooking recipes. In the last couple of months, the platform has signed deals with ChatGPT maker OpenAI and Google to share its user data, and now the company is reportedly gearing up to enter the search engine space. In its latest note to investors, Reddit CEO Steve Huffman said that the company is now focusing on 'areas that will drive results for our most pressing needs', one of which involves 'making Reddit a go-to search engine.' Hoffman went on to say that Reddit is one of the few platforms on the internet that is 'positioned to become a true search destination' as it has knowledge and conversations of real people, something which cannot be found elsewhere. The platform, which has more than 70 million weekly users, is now looking to make search a 'central feature.' To achieve this, the company says it will be making Reddit Answers, a feature currently available in more than a dozen countries to everyone and will make the search box more visible. Launched in December 2023, Reddit Answers is an AI-powered feature that has grown from 1 million users in 2024 to 6 million weekly users in the last year. In a statement to TechCrunch, CEO Steve Huffman said, 'We're unifying those [Reddit search and Reddit Answers] into a single search experience, and we're gonna bring that front and center in the app. So, whether you're a new user opening the app for the first time or a returning user opening the app, that search box will be present immediately.' With low-quality content taking over the internet in the last few years, people around the world have been adding the word 'Reddit' at the end of their search query to get relevant results. But with Google now offering its very own AI mode that automatically summarises content from all over the internet, Reddit to looking to jump into action by offering a similar product.
