ISMPP EU 2023
Will AI tools change the way we make enhanced publication content?
As has been the case over the past several years, numerous presentations and posters looked at plain language summaries and other publication enhancements. At this meeting, there were also suggestive studies exploring the potential role of AI tools in this area. Here are a few of our highlights.
McMinn et al. compared plain language summaries generated by humans with those generated by natural language processing AI tools.1 The AI showed promise but had consistent areas of weakness, such as accurate reporting of results, and would require human input to be fit for purpose. This is a rapidly evolving field: the recent release of ChatGPT prompted the World Association of Medical Editors to release a guidance statement,2 and it is likely that other organizations will follow suit.
Beeby et al. examined the quality of plain language summaries translated by browser-based software.3 In a small sample size, the quality of translations was promising - increased use of this approach could widen access to research for non–English-language speakers, including patients.
OPEN research: Discoverability of enhanced publication content
Literature searches by Black et al. confirmed that enhanced content formats are difficult to locate, and often incorrectly indexed even when presented as standalone peer-reviewed items.4 Publisher interviews suggested approaches to improving this situation.
AI, artificial intelligence.
1. McMinn D et al. Presented at 2023 European Meeting of ISMPP; London, UK. Poster #21; 2. World Association of Medical Editors. https://wame.org/page3.php?id=106 (accessed February 2023);
3. Beeby C et al. Presented at 2023 European Meeting of ISMPP; London, UK. Poster #5; 4. Black S et al. Presented at 2023 European Meeting of ISMPP; London, UK. Poster #29.