What we learned from reviewing AI tools for news
Jinfo Blog
10th July 2025
Abstract
Our article, “How should information managers respond to the call to use free AI tools for news?” and report, “Can you rely on open-source AI for news? A review of Perplexity and ChatGPT” contain our detailed findings, and this blog reflects on some of our findings.
Item
Much has been said about the intersection of news and AI, and Jinfo took a closer look at this evolving topic during Spring 2025 in our article "How should information managers respond to the call to use free AI tools for news?" and report "Can you rely on open-source AI for news? A review of Perplexity and ChatGPT".
These explore how GenAI tools provided by the paid aggregators (like Factiva) compared to what the “free” AI platforms (like ChatGPT and Perplexity) can provide.
Free AI platforms are becoming increasingly sophisticated, offering real-time search, summarisation, and research capabilities. In response, traditional paid news aggregators have introduced AI-driven summarisation tools.
As free AI platforms continue to improve and broaden their source coverage, traditional paid aggregators face growing pressure and risk losing momentum. As one experienced user told Jinfo: "This could replace our premium news provider."
During our research, two issues emerged on both sides:
- The role of sources
Traditional paid aggregators typically draw on solid, reputable sources, the Financial Times being a prime example. Their focus is on reliability, carefully selecting from a broad range of credible publishers (especially from news aggregators like Factiva) and enabling users to narrow searches to specific sources.
Free GenAI platforms, by contrast, rely mainly on publicly available content. That’s starting to change, though, with many publishers signing deals with platforms such as Perplexity and ChatGPT, opening access to parts of their archives.
Despite these advances, a significant portion of news content remains out of reach. Plus, free tools generally place few restrictions on how their output can be used.
But, paid providers operate under strict licensing agreements with their contributors. Many sources within paid aggregators have yet to be cleared for AI-driven applications, which narrows the scope of their AI features. -
The technology challenge
GenAI tools from traditional paid aggregators are less sophisticated than free AI platforms.
The latter leverage advanced large language models that deliver dialogue-based, detailed answers, making them feel more natural and useful.
Whereas, aggregators' solutions often produce shorter summaries, with limited functionality, which can frustrate users seeking deeper analysis.
The landscape of AI-driven news summarisation is evolving quickly. Free AI platforms lead in terms of technology and ease of use, while paid aggregators offer more controlled and reliable content. Each has clear strengths and drawbacks.
Ultimately, no platform offers a perfect solution. Whether using a premium paid aggregator or a free AI platform, it's essential to be mindful of the limitations, paying close attention to quality, relevance, and the source of any summaries or insights before putting them to use.
At the end of the day, the ideal solution would be to combine the two, allowing the best sources to benefit from the best technology. As Factiva clears more and more sources for AI use, this is likely to become possible at some point in the future.
For more insights, see the Jinfo article "How should information managers respond to the call to use free AI tools for news?" and the Jinfo report "Can you rely on open-source AI for news? A review of Perplexity and ChatGPT".
- Blog post title: What we learned from reviewing AI tools for news
- Link to this page
- View printable version
Register for our next Community session:
![]()
Transforming knowledge management at BASF – GenAI and the evolution of QKnows
10th December 2025
Latest on our YouTube channel:![]()
Read on the Blog:
December 2025 update
3rd December 2025
- Jinfo wins CILIP’s inaugural “McFarlane & Ward Information Management” award
4th December 2025 - December 2025 update
3rd December 2025 - Review of Matchplat – combining AI with traditional industry code searching
27th November 2025
- Team roles and AI (Community) 26th February 2026
- Team demand and AI (Community) 22nd January 2026
- Transforming knowledge management at BASF – GenAI and the evolution of QKnows (Community) 10th December 2025