QED Science Launches Insights Feed to Filter Scientific Literature
New tool aims to help researchers evaluate paper validity rather than just finding more articles to read.
Key points
- QED Science launched Insights Feed this week for life-science researchers.
- The tool addresses the difficulty of determining which science is reliable.
- AI discovery tools often fail to distinguish between shaky preprints and rigorous studies.
QED Science launched Insights Feed this week, a personalized discovery tool designed for life-science researchers. The product aims to solve the problem of deciding which scientific papers are worth reading and whether the underlying science holds up, rather than simply generating longer reading lists.
Niv Mastboim, CEO of QED Science, stated that researchers do not need another tool that provides more papers to read. Instead, they require a better method to decide which science they can build upon. The launch comes as AI agents can read dozens of scientific papers in minutes, yet researchers continue to struggle with keeping up due to the volume of available literature.
Context and Background
The volume of scientific literature has increased significantly, with the number of articles indexed in major research databases rising by roughly 47 percent between 2016 and 2022, according to The Next Web. Automating the search process without improving judgment simply moves the bottleneck from finding papers to sorting through them for relevance. Most discovery tools, whether AI-powered or not, start from a query, which can surface work that mentions the right protein or disease but does not bear on a scientist’s actual question.
AI agents inherit blind spots from their prompts and summarize everything with similar confidence. This means a shaky preprint and a rigorous study can return looking almost identical, failing to help researchers distinguish between high-quality science and less reliable work. The Next Web reported that AI discovery tools often do not look at the validity aspect at all, leaving the hard part of the job—evaluating relevance and quality—unsolved.
Who Is Affected
Life-science researchers who rely on current literature are the primary users of this new tool. For more than a decade, scientists used Science Twitter as a filter, debating preprints and finding collaborators. Mastboim noted that Science Twitter used to be like the best part of conferences, acting as a social filter for new developments. However, a 2025 survey by David Shiffman and Julia Wester found that two-thirds of scientists said Twitter had become much less useful for learning about new developments in their field.
Many scientists have since moved to Bluesky, but social feeds still depend on who users follow and what they share. Agents, by contrast, return only what users already knew to ask for. The erosion of the social filtering mechanism on Twitter has left a gap in how researchers discover and validate new scientific work, which QED Science aims to fill with its validity engine.
What should you do?
Researchers should evaluate whether current AI discovery tools are providing relevant results or just increasing the volume of unread papers. Institutions may want to assess how their staff are filtering scientific literature and whether new tools like Insights Feed could reduce the time spent sorting through irrelevant studies. Monitoring the shift in scientific communication platforms is also important, as social filters like Twitter have changed.
What to do and how to stay safe: QED Science
- Review current literature search workflows to identify bottlenecks in sorting and evaluating paper relevance.
- Test new discovery tools to see if they improve judgment on paper validity rather than just increasing search volume.
- Monitor changes in scientific communication platforms that may affect how researchers discover and validate new studies.
- Assess whether AI agents are distinguishing between rigorous studies and shaky preprints in your specific field.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What problem does QED Science's Insights Feed solve?
It helps life-science researchers decide which papers are worth reading and whether the science behind them holds up, addressing the issue of excessive volume and poor filtering.
How has the volume of scientific literature changed recently?
The number of articles indexed in major research databases rose by roughly 47 percent between 2016 and 2022, according to The Next Web.
Why did scientists find Twitter less useful for research?
A 2025 survey found that two-thirds of scientists said Twitter had become much less useful for learning about new developments in their field, leading many to move to other platforms.
