This is how open data and AI could boost the impact of scientific research

At this moment, keystrokes are capturing decades of scientific knowledge that culminated in a paper, or insights from an experiment recorded in an electronic laboratory notebook. A repository receives an article that has the key to a breakthrough buried in its text. This moment is being repeated around the globe every day of every year

The knowledge needed for the next great scientific breakthrough may have already been written, but it may go undiscovered by tools that view knowledge as singular items to merely be found instead of parts of a collective intelligence.

Global expenditure on R&D exceeds $2 trillion annually. In 2020, the US led all countries, spending $609.7 billion, and Asia led all regions with $1.07 trillion in spending on R&D. Recently, US President Joe Biden proposed a $250 billion investment in research, and expressed a desire to increase spending on research to 2% of US GDP.

Record-breaking development of COVID-19 vaccines highlight the benefits of increased investment in R&D. The global pandemic has spawned unprecedented collaboration in scientific communities around the globe, and sparked the rapid evolution of technologies to combat the virus, including Artificial Intelligence.

The Third Industrial Revolution paved the way for the digital dissemination of the products of R&D; structured and unstructured scientific big data, such as datasets and journal articles. In the Fourth Industrial Revolution, AI has advanced drug and genomic discovery by analyzing the proliferation of scientific data facilitated by the Third.

As COVID-19 spread in communities around the world, publishers made what might otherwise be paywalled research available as Open Access via initiatives such as the COVID-19 Open Research Dataset (CORD-19). Scientists were invited to not only access the research, but to leverage AI to expedite discovery across the collection, which has grown exponentially, with peaks of more than 5,000 articles being produced per week.

As a result, several search engines were built for COVID-19 research. However, these tools only improved the ability to find documents in the collection, which had no impact on accelerating an understanding of the research itself. Still more tools examined and mapped topics in the research, creating visualizations and navigation aids.

Such projects, while scratching the surface of what is possible with AI-democratized scientific knowledge, largely maintain a perspective of knowledge in R&D as linear. Research is conducted, a paper is written, and that paper is stored in a repository to be searched after being enriched with keywords.

The circularity of scientific knowledge as digital assets is best displayed by companies like Benevolent AI, who recommended drugs for treatment of COVID-19 by analyzing assets, including literature. In this case, research was viewed as a collective in order to return derived insight to the scientific community. In turn, scientists were empowered to initiate clinical trials, and produce new knowledge that would become digital assets.

In a linear perspective of digital assets in R&D, knowledge is downgraded to disconnected, individualized information. In a circular perspective, knowledge is not being underutilized or wasted, but recompiled and re-used to create more knowledge that has been derived from the collective.

However, the technology leveraged by AI-powered drug discovery companies is largely out of reach to millions of scientific professionals around the world. High costs and scope that is either too broad or too narrow drive inaccessibility. If the true promise of AI for R&D is a continuous feedback loop of knowledge creation and consumption, then it must be democratized to maximize impact.

Circular innovation is optimized as each researcher and digital asset is added to an AI-enabled ecosystem that analyzes collective intelligence. By removing the barriers of scope and costs, more researchers have access to not only consume knowledge, but to create new knowledge as byproducts of what AI uncovers.

By liberating scientific knowledge for analysis with AI, and democratizing access to the insights that are uncovered, the impact of R&D is increased exponentially. Uncovering the cache of knowledge that will unlock the next cure is made possible, and placed in the hands of the many, not the few.