The academic world has a readability problem. While a journal article or research paper captures key information well, it’s time consuming for experts to consume, and complex jargon and prose can make it inaccessible to the general public.
A recent analysis by The Economist revealed a concerning trend: academic writing has become increasingly difficult to read over the past 80 years, with the most dramatic decline in readability occurring in the humanities and social sciences.
Why is Research so Hard to Read?
What’s making it so difficult to navigate academic research?
- Dense technical vocabulary that requires specialized knowledge
- Complex sentence structures that challenge even experienced readers
- Information-heavy content that can overwhelm readers
- Writing that prioritizes technical precision over clarity
- The "curse of knowledge" - authors unconsciously assuming readers share their deep subject understanding
These barriers limit the broader impact of valuable research by making it inaccessible to students, practitioners, policymakers, and interested members of the public who could benefit from these insights.
AI as a Bridge Between Academia and Broader Audiences
Artificial Intelligence, particularly Large Language Models (LLMs), offers a promising solution to this accessibility challenge. AI can transform dense academic content into various formats tailored to different audiences' needs and expertise levels. Here's how:
1. Multi-Level Summaries
AI can generate different versions of the same content for various audiences:
- "Lay summaries" that make research accessible to general readers
- Expert summaries that help specialists quickly grasp key findings
- Student-friendly versions that support learning at different educational levels
2. Interactive Learning Tools
AI can transform static papers into dynamic learning resources by:
- Generating thought-provoking questions that encourage deeper engagement
- Creating research highlights suitable for presentations
- Identifying key concepts and providing additional context where needed
3. Research Innovation Support
For early-career researchers, AI can:
- Help generate new research hypotheses based on existing work
- Identify potential research directions
- Connect findings across different papers and fields
A Real-World Example: Making Astronomy Accessible
In the Bright Future of AI report that Hum released earlier this year, we shared an example that combines content data and generative AI at scale to generate other useful media.
Let's look at how AI can transform complex astronomical research into valuable content for different audiences.
Taking a paper about the IM Lup protoplanetary disk study from UVA astronomer Ilse Cleeves and team, AI can generate new content for a variety of audiences.
For General Readers (9th Grade Level):
A clear explanation of how scientists are studying the building blocks of planets around young stars, using everyday language and relatable comparisons. This makes cutting-edge research accessible to curious minds without sacrificing accuracy.
Overview: This research focused on understanding the amounts and types of gases like carbon (C), nitrogen (N), and oxygen (O) in a young planetary system called the IM Lup protoplanetary disk. They used observations from a powerful telescope called ALMA.
Why This Matters: Protoplanetary disks are clouds of gas and dust around young stars, where planets start to form. By studying these disks, scientists can learn about the materials that eventually make up planets. This particular study looked at the gases in one of these disks, which is important because the gas can end up on planets as their atmosphere.
Key Discoveries:
- Gas Ratios: They found that the ratio of carbon to oxygen (C/O) in the gas is about 0.8, which is higher than the sun's ratio of about 0.54. This means there's more carbon compared to oxygen in this young planetary system than in our sun.
- Water Ice and Carbon: A lot of the oxygen is locked up in water ice and larger dust particles that settle in the middle of the disk. Some of the carbon is also trapped in less volatile (less easily vaporized) molecules.
- Nitrogen: Nitrogen, on the other hand, is mostly found as a gas (likely N2) and doesn't get trapped in ice or dust as much as oxygen and carbon.
What This Means for Planets: Understanding these gas ratios is crucial because they can influence the type of atmospheres that planets develop as they form in these disks. For instance, having more carbon than oxygen could make the atmospheres of new planets very different from Earth's.
The Bigger Picture: These findings might help predict the composition of gases around other young stars and the atmospheres of planets forming there. It's like getting a sneak peek into how different planets might be born and what their air could be like, which is super important for understanding the variety of planets in our universe.
For Experts:
A concentrated summary focusing on methodologies, key findings, and implications, helping researchers quickly determine if the paper is relevant to their work.
Employing ALMA observations and chemical modeling, this study on the IM Lup protoplanetary disk focuses on gas-phase C, N, and O abundances. It finds a super-solar C/O ratio and a significant N/O ratio variation in the molecular layer, particularly evident between radii of 20 and 300 au in a pseudo-steady state. The transformation of H2O ice to gas-phase atomic oxygen and nitrogen’s stability in N2 form are key factors affecting these ratios, crucial for understanding planet formation processes.
For Students and Educators:
Discussion questions, learning materials that bridge the gap between complex research and classroom learning, and suggestions for future research.
Future Hypotheses:
"Differential Processing of Volatiles in Protoplanetary Disks and Its Effect on Planet Formation"
Rationale: Building on the findings from the IM Lup disk study, which uncovered a super-solar C/O ratio and distinctive processing of nitrogen, this hypothesis proposes that the differential processing of volatiles like carbon, oxygen, and nitrogen in protoplanetary disks plays a crucial role in determining the chemical composition and potential habitability of forming exoplanets.
Potential Future Research Directions:
Comparative Studies: Investigate the volatile compositions of other protoplanetary disks using similar methodologies to understand if the IM Lup disk's characteristics are common or anomalous.
Modeling Variations in Volatile Processing: Develop and test models that simulate how different levels of volatile processing impact planet formation and the resulting atmospheric compositions.
Link to Exoplanet Atmospheres: Explore the potential correlations between the volatile compositions of protoplanetary disks and the observed atmospheric compositions of exoplanets to validate the hypothesis.
The Future of Academic Communication
As AI tools continue to evolve, we can envision a future where academic research automatically comes with a suite of companion materials tailored to different audiences. This transformation could:
- Increase research impact by making findings accessible to more stakeholders
- Improve science communication and public understanding
- Accelerate innovation by making it easier to build upon existing research
- Support education at all levels by providing appropriate entry points to complex topics
Making academic research more readable is all about maximizing the societal impact of research. AI offers us tools to bridge the gap between specialized academic knowledge and the broader world, ensuring that valuable insights can reach and benefit everyone who needs them.
Scaling Content Intelligence for Broader Impact
Could you plug any paper into OpenAI, prompt it to generate a lay summary, and get a solid answer? Absolutely!
Individual paper transformations are valuable, but to run every single paper individually through a generative AI system this way would be quite expensive in time and money.
To scale a solution like this, you need to be able to process and interconnect entire bodies of research systematically and accurately. Building effective content intelligence requires robust infrastructure that can:
- Accurately parse and understand academic text
- Maintain consistent categorization across diverse content
- Track relationships between concepts and papers
- Generate appropriate outputs for different audience needs
This is where sophisticated AI engines like Hum's Alchemist come into play! By understanding not just the words but the concepts, relationships, and implications within academic content, publishers can generate more valuable and contextually appropriate outputs for different audiences.
Learn more about how Hum is helping publishers boost content discoverability with advanced taxonomy development and tagging, AI-driven recommendations, and data intelligence, ensuring high-value content reaches the right audiences.