AIDA for Knowledge Extraction – Research Review on an Unprecedented Scale
Juhani Teeriniemi, Lead developer
Ai4Value flagship product AIDA – AI Document Assistant has finally made its way into the academic world. In a recent article published in iScience, AIDA was used to investigate a simple question: “Is there progress in the development of adsorbents for water treatment?”
It all began one Friday afternoon when Tero Luukkonen from the University of Oulu gave us a call. He had just finished reading yet another paper on adsorption in water treatment and wondered: How is it possible that thousands of studies are published each year, yet only a handful of adsorbents are used in commercial water treatment plants? Are genuinely novel materials being developed – and if so, what are they? Maybe AI could help. It could read all published research and summarize the progress made in the field.
AIDA was indeed the right tool for the job. But first, we had to get more precise about what we meant by “progress.” We chose to focus on adsorption capacity for commonly used benchmark adsorbates – such as methylene blue.
We used AIDA to tabulate data from 11,664 academic articles, organizing them into four columns: adsorbent name, adsorbate name, capacity, and adsorbent category. With this dataset, we could analyse which categories of adsorbents showed the strongest development over decades of research.
The results were striking. Majority of studies contributed little advancement when using capacity as the key metric. However, certain engineered nanomaterials – carbon nanotubes, MXenes, and metal-organic frameworks – demonstrated strong progress.
We suspect that several non-technical barriers are holding back the adoption of these new materials in real-world water treatment. Any new technology in this field must be necessary, cost-effective, and low-risk.
We were excited to see that data extraction was practically flawless. In a manual verification, 98% of the results were identical, and in the remaining 2%, the differences were a matter of interpretation.
One thing seems clear: AI-powered document assistants like AIDA have much to give in academic research. This powerful research tool can be tailored to your company’s needs.