SciencePOD

Artificial intelligence tools are vital to open access dissemination

Hey, industry leader! Hey, potential collaborator! Funders! Look at what I’ve published! Waiting passively to appear in online searches and feeds is not enough. Scientists must be active and make some noise about their work – more noise than everyone else. Simple summaries of findings are among the best ways to amplify exposure and get noticed – and it is now becoming much easier thanks to recent advancements in Artificial intelligence (AI)-based summarisation engines. Automated summaries and content help scientists and publishers to get noticed on social networks and through more traditional digital distribution channels. 

All researchers know it is difficult to get noticed and the statistics on Open Access (OA) show why. The number of published papers alone – excluding pre-prints – has climbed exponentially to 2.5 million per year, according to estimates by STM Association, a scholarly publishing trade organisation. Perhaps even more significant is the rise in OA papers, which in 2021 represented 39% of the total number of publications in volume (Delta Think). A study by Heather Piwowar and colleagues from Our Research, forecasts that OA papers will represent 70% of all article viewsby 2025. 

Open access upward trend

Whilst OA is great for inclusive knowledge sharing – in theory – it poses significant challenges for academic publishers and scientists alike. Most top-ranking journals still tend to be ‘closed’ (you need an institutional subscription to access them), or they are ‘pay as you go’ for the occasional download. And publishers charge authors high fees to make papers freely accessible as so-called ‘gold OA’. Scientists must decide a) if their work merits a top journal; and b) if they or their institution can afford the extra fees. Some studies suggest that OA papers have a citation advantage; for example, a recent analysis found that two-thirds of older studies (2001-2015) showed an OA advantage, rising to 82% for more recently published work (2016-2021). 

Institutions have overcome the cost barrier to OA by creating repositories of open access versions of papers. Known as ‘green OA’, they are freely accessible, but cumbersome to use; readers must often dig deep into an author’s institutional website to find where green OA papers are stored. Networking and aggregation sites like ResearchGate and CORE offer alternative routes to find, download or request free copies of papers. 

Discover great science using artificial intelligence

But perhaps the biggest challenge that publishers and academics face is the dilution effect. How can one paper ever reach the surface and get noticed when it floats in such a vast ocean of published (and increasingly accessible) work?

Unsurprisingly, artificial intelligence is starting to prove its worth. Today, behind the vast army of real-life authors, editors, reviewers and journal publishing teams, AI tools are helping papers to make their mark, seemingly against all the odds.

Let’s assume that a paper has passed its many qualitycontrol and publishing hurdles. The science is sound, reviewers are satisfied, a new open access paper is born. How can AI help these articles get the attention they so deserve?

First, there are many tools to improve a paper’s visibility and discoverability, for example, on search engines such as Google Scholar, specialist academic systems like Scopus, or through engines like Semantic Scholar. Platforms like iris.ai help users to monitor all the latest developments in a given field. So, it is important that authors get the basics right, not just for human readability but for search engines too. A paper’s title, abstract, keywords and metadata are all vital components, but usually beyond the publisher’s control once a manuscript is submitted. 

Several companies provide services that assist publishers with indexing and distribution. The ScienceOPENrepository enables publishers to promote and curate content, for example using banner marketing and search highlights. The platform helps to build researcher networks and communities of interest around specific topics. Cadmore Media, meanwhile, can help scientists get noticed using metadata-rich media streams that help others discover audio-visual recordings of congress and conference presentations and content. 

Publishers should also be aware of new platforms that may make these traditional search methods and curated platforms redundant. Juisci, for example, uses AI to process content from top journals automatically; the platform uses a recommendation algorithm to deliver a ‘live feed’ of relevant papers to a user based on their profile, scientific specialisms and behaviours. R Discovery is similar and pushes out highlights from the three most relevant papers each day, tailored to each unique user. AI can create uniquely personalised rabbit holes for users to explore.

Clarify the message

Even when a paper pops up on your screen, what makes your read and click? Users are time poor, and may not even read a whole abstract. Artificial intelligence can help here too, by summarising a paper’s content in language that is accessible and clear, with sufficient information to understand the context of a study as well. Automated summarisation is finally helping OA papers to achieve the goal of reaching wider audiences and help stimulate cross-fertilisation between fields.

In its mission to communicate standout science, SciencePOD has developed its specialist SUMMSci summarisation engine through a research collaboration with Avignon Université. SUMMSci can summarise thousands of papers in an instant. It generates short overviews in plain English, so readers (researchers, R&D industry, media organisations, science and medical publishers) can quickly spot relevant work or hidden gems. Summaries include key methodological detail, results and discussion points. 

SUMMSci is also behind SciencePOD’s ScioWire feed. This takes the advantages of instant summarisation to the next level, allowing users to create personal feeds that summarise all the latest OA research, based on their own keyword preferences.

Promote and amplify

Science is about sharing ideas and lively, open debate. It is arguably one of the most socially active professions on the planet! Perhaps the next frontier for Artificial intelligence in scientific publishing is in the guise of a virtual influencer. These artificially intelligent personas have their own social media accounts, TikTok channels and a huge global fan base. Could a science publisher dare to use such sophisticated creations to promote top papers to a global following? Headlines guaranteed!

That said, it is important not to get carried away by the AI hype. Certainly, these tools can help researchers, publishers and reporters in their efforts to identify and share the very latest discoveries. But nothing quite hits the mark like a powerful story. 

A story draws you it, it stirs emotions, it engages the mind and…the soul. You are intrigued, you remember, you want more, you click. While AI offers the promise of speed and productivity, we still need professional science and medical communicators to tell stories that resonate with real people in the real world.

Discover the ScioWire research newsfeed: summarised scientific knowledge ready to digest.

X