Poll finds public turning to AI bots for news updates
PARIS: People are increasingly turning to generative artificial intelligence chatbots like ChatGPT to follow day-to-day news, a respected media report published Tuesday found.
The yearly survey from the Reuters Institute for the Study of Journalism found 'for the first time' that significant numbers of people were using chatbots to get headlines and updates, director Mitali Mukherjee wrote.
Attached to Britain's Oxford University, the Reuters Institute annual report is seen as unmissable for people following the evolution of media.
Just seven percent of people report using AI to find news, according to the poll of 97,000 people in 48 countries, carried out by YouGov.
But the proportion is higher among the young, at 12 percent of under-35s and 15 percent of under-25s.
The biggest-name chatbot -- OpenAI's ChatGPT -- is the most widely used, followed by Google's Gemini and Meta's Llama.
Respondents appreciated relevant, personalised news from chatbots.
Many more used AI to summarise (27 percent), translate (24 percent) or recommend (21 percent) articles, while almost one in five asked questions about current events.
Distrust remains, with those polled on balance saying AI risked making the news less transparent, less accurate and less trustworthy.
Rather than being programmed, today's powerful AI 'large language models' (LLMs) are 'trained' on vast quantities of data from the web and other sources -- including news media like text articles or video reports.
Once trained, they are able to generate text and images in response to users' natural-language queries.
But they present problems including 'hallucinations' -- the term used when AI invents information that fits patterns in their training data but is not true.
Scenting a chance at revenue in a long-squeezed market, some news organisations have struck deals to share their content with developers of AI models.
Agence France-Presse (AFP) allows the platform of French AI firm Mistral to access its archive of news stories going back decades.
Other media have launched copyright cases against AI makers over alleged illegal use of their content, for example the New York Times against ChatGPT developer OpenAI.
- News still a draw for X -
The Reuters Institute report also pointed to traditional media -- TV, radio, newspapers and news sites -- losing ground to social networks and video-sharing platforms.
Almost half of 18-24-year-olds report that social media like TikTok is their main source of news, especially in emerging countries like India, Brazil, Indonesia and Thailand.
The institute found that many are still using Elon Musk-owned social media platform X for news, despite a rightward shift since the world's richest man took it over.
'Many more right-leaning people, notably young men, have flocked to the network, while some progressive audiences have left or are using it less frequently,' the authors wrote.
Some 23 percent of people in the United States reported using X for news, up eight percent on 2024's survey, with usage also rising in countries like Australia and Poland.
By contrast, 'rival networks like Threads, Bluesky and Mastodon are making little impact globally, with reach of two percent or less for news', the Reuters Institute found.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Star
an hour ago
- The Star
‘Information is speed': Nascar teams use AI to find winning edges
CONCORD: Margins in Nascar have never been smaller. Whether it's the leveling effect of the Next Gen car or the evolving technological arms race among teams, the Cup Series has never been tighter. And as parity grows, so does the need to uncover even the slightest competitive advantage. That's where artificial intelligence comes in. From performance analysis to data visualisations, AI is playing an increasingly pivotal role in how race teams operate across the Nascar garage. Teams are using AI not just to crunch numbers, but also to make quicker decisions, generate strategic insights – and even rewrite the way they approach race weekends. 'It just builds a little bit more each year,' said Josh Sell, RFK Racing's competition director. 'We're doing more now than we were a year ago. And we'll probably be doing more a year from now than we are sitting here right now. It just continues to evolve.' Asking better questions, getting smarter answers The rise of AI in Nascar mirrors the broader tech world. Early large language models – or LLMs – were trained to answer basic questions. But now, they can cite sources, detect tone and reason through complex decisions. That opens up a new world for how teams evaluate everything from strategy calls to post-race feedback. For example, a full race's worth of driver and crew radio chatter can be fed into an AI model that not only identifies which calls worked and which didn't, but also interprets tone and urgency in real time. 'Information is speed in this game nowadays,' said Tom Gray, technical director at Hendrick Motorsports. 'He who can distill the information quicker and get to the decision quicker, ultimately, is going to have the race win. If you can control the race or make that decision that gets you in control of the race at the end, you're going to be win the one who wins.' Finding the time where it matters AI is also helping teams develop talent and streamline operations. Even if someone on the team isn't an expert in a particular field, AI can help them learn new skills faster. That's especially important in the highly specialised Cup Series garage – and it could help smaller teams close the gap with bigger operations. RFK Racing, now a three-car Cup Series team, is already seeing those benefits. AI helps reduce the hours team members spend manually analysing photos or videos. Instead of having a crew chief sort through everything, the software flags the most relevant material and delivers it quickly. On the technical side, the team is also using tools like ChatGPT to assist with software development, solving coding problems in various languages and freeing up engineers to focus on execution. 'It's trying to figure out ways where, instead of having a crew chief spending three hours studying whatever it might be – photos, videos – if we can shorten that to an hour of really impactful time,' Sell said. 'Looking at things that are important to them, not searching to find those things. That's the biggest gain we see, and certainly whether it's through the week or on race weekends, time is our limiting factor. 'You have a finite amount of time from the time practice ends to when the race starts. What you're able to do to maximise the efficiency of that time is kind of a race in and of itself.' Visuals, velocity and vintage data At Hendrick Motorsports, the winningest team in Cup Series history, AI is being used both to look ahead and to look back. The team now works closely with Amazon Web Services (AWS) – a relationship that began after Prime Video sponsored one of its cars. The partnership has accelerated Hendrick's use of AI across several key areas. One of those is visual communication. Engineers are now generating images to help share ideas, whether they're pitching a new part or breaking down a technical strategy. That ability to visualise complex concepts instantly helps everyone stay aligned and efficient. Hendrick is also leveraging its four decades of data. The team can now go back and test old strategies, setups and decisions using AI to predict how past insights might inform future success. 'We've had a long history in the sport,' Gray said. 'Not only can we look forward, but we can also look backward, back-test all the information we have, and see how that predicts the future.' – The Charlotte Observer/Tribune News Service
The Star
6 hours ago
- The Star
Tesla invites select few to Texas robotaxi trial with front seat safety monitors
Arthur Maltin, a test driver with The Dawn Project, performs a safety test on a self-driving feature in a Tesla Model Y during a protest against Tesla robotaxis, ahead of the Tesla robotaxis' official services in Austin, Texas, U.S., June 12, 2025. REUTERS/Joel Angel Juarez (Reuters) -Tesla has sent out invitations to a small group of people to join the limited test of its robotaxi service in Austin, Texas, which is tentatively set to start on Sunday, according to posts and email screenshots on social media. The invitations said a Tesla employee will accompany riders in the front passenger seat, the posts showed. A successful trial in Austin will be crucial for Tesla, as car buyers shun the company following its CEO Elon Musk's embrace of hard-right politics in the U.S. and worldwide. The company has shifted from building a cheaper EV platform to robotics and artificial intelligence. Much of the company's valuation rests on that bet. Musk has said he will focus on safety in the trial, with humans remotely monitoring the vehicles. He has said he expects to be able to scale the robotaxi quickly. Tesla plans to deploy around 10 Model Y SUVs that run on a version of Tesla's full self-driving driver assistance software, Musk has said. The start could be delayed and the service might be limited or unavailable in bad weather. Riders have to be aged 18 or above to take a ride. "Tesla is rolling out the Robotaxi service extremely cautiously, which is good. Baby steps," posted Omar Qazi, an user with 635,100 followers who posts often about Tesla using the handle @WholeMarsBlog and received an invite. "Very much so," CEO Elon Musk responded. But commercializing autonomous vehicles is an expensive, risky process. Tesla and others such as Alphabet's Waymo and Amazon's Zoox have faced federal investigations and recalls following collisions. Industry experts have questioned the efficacy of Tesla's self-driving technology that depends mostly on cameras and AI, without redundant sensors such as lidar and radar, claiming fog, heavy rain, and glaring sunlight can hamper safety. On Wednesday, a group of Democratic lawmakers from the Austin area called on the company to delay the rollout until September, when a new state law governing autonomous vehicles takes effect. Users who confirm their presence in Austin can download Tesla's Robotaxi app, which they use to call a vehicle. "Through this exclusive preview, you'll have the opportunity to provide valuable feedback on our Robotaxi service," one of the screenshots could not immediately verify the authenticity of the screenshots. (Reporting by Akash Sriram in Bengaluru and Abhirup Roy in San Francisco, Editing by Peter Henderson and Louise Heavens)
Malaysian Reserve
7 hours ago
- Malaysian Reserve
Artificial Intelligence (A.I.) in Life Science: New Survey of 408 Researchers Reveals Split Sentiment, Surging Adoption, and Rising Trust Barriers
ARLINGTON, Va., June 20, 2025 /PRNewswire/ — New survey data from 408 global scientists and researchers reveals a rapid increase in AI adoption—alongside mounting hesitation around trust, training, and usability. As budgets tighten and workflow efficiency becomes more critical, scientists are both accelerating the use of AI and raising red flags about its limitations. The latest findings from BioInformatics' Beyond the Bench series uncover how academic, government, and industry researchers are evaluating and integrating AI into their tasks and workflows. While 87% of professionals report using AI for work-related research tasks—a sharp increase from 75% in 2023—the survey also exposes rising concerns around data fidelity, cybersecurity risks, and the readiness of current AI solutions for complex life science environments. This snapshot of market sentiment arrives at a pivotal moment, as companies seek faster insights and operational efficiencies but face institutional barriers to confident AI implementation. Key Insights from the AI in Life Sciences Survey: AI adoption is high, but value varies: 87% of researchers report using AI in their workflows, yet only 27% of avid users say it brings high value. Top providers gaining traction: Scientists identified Microsoft, Thermo Fisher Scientific, and Google DeepMind as leading organizations in developing or deploying AI in life sciences. Efficiency drives demand: Most users cite faster processing, improved throughput, and time savings as key benefits. Training and infrastructure gaps persist: Lack of regulatory clarity, insufficient user training, and organizational readiness are major hurdles to adoption. Click here to download Perceptions of AI in Life Science Survey Data 'As life science organizations look to AI to drive productivity, these findings show that value is conditional,' said Richa Singh, VP of Market Insights at BioInformatics. 'Vendors must address usability, trust, and real-world application if they want adoption to translate into impact.' About Beyond the Bench Beyond the Bench is a free monthly intelligence series created by BioInformatics to help life science and diagnostics companies keep pace with shifting customer sentiment and commercial priorities. Powered by the Science Advisory Board—BioInformatic's proprietary network of more than 55,000 qualified life science professionals—each report delivers timely, survey-based insights into the market dynamics shaping product strategy, messaging, and customer engagement. Click here to subscribe. Access & Get Involved Download the AI in Life Science Free Data Highlight – Explore how your customers perceive AI, what they need, and where adoption is stalling. Inside the Minds of Scientists: How AI is Changing Life Science Research: Access the webinar on demand & download the full slide deck for deeper insights. Join the Science Advisory Board – Participate in future surveys and shape the future of life science research. About BioInformatics BioInformatics, part of the Science and Medicine Group, is a leading market research and advisory firm serving the life science and diagnostic industries. The company delivers custom and syndicated research powered by a proprietary global panel of more than 55,000 professionals. With deep domain expertise and real-time market access, BioInformatics enables commercial teams to make better, faster, and more strategic decisions.
