Is Artificial Intelligence (AI) the future for scientific content creation? 

Artificial Intelligence (AI)– the hot topic right? AI is something we’ve had many conversations about at HRS Communications. As the hype grows and also the concerns around the potential impact that AI may have on content creators, we challenged one of our students; Freya, with the task of taking a deeper look into this constantly evolving technology as part of her Uni research project. 

During her time at HRS, Freya began investigating the effectiveness of AI in creating credible nutritional content compared to human content creation. AI is a system which works on machine learning, i.e. pulling information from a range of different web sources to analyse and interpret large amounts of data and identify trends to create relevant content. This fascinating system has been gaining traction over the years due to its efficient nature which sparked the idea of conducting a pilot case study within this area. Curating credible content can be a time-consuming task which demands creativity and scanning numerous journal articles to ensure the content is evidence-based. Therefore, the aim of this study was to assess whether AI is a viable option for HRS’s clients and other healthcare companies in increasing the efficiency and productivity of content creation.  

When conducting the research, AI and Freya were given a set of criteria, by HRS, to write a blog about ‘Nutrition in Pregnancy: UK Government Guidelines and Recommendations’ with an informative and engaging tone of voice, aimed at Dietitians. Fascinatingly, the AI blog took less than 5 minutes to create the blog. However, the output of the AI blog was far from HRS’ usual high standards.  

Assessors analysing the two blogs stated the AI blog had no cohesion or flow, contradicted its facts and was not referenced. This raises concerns around ethical issues as if a Dietitian were to apply some of AI’s recommendations into practice, it may cause patient harm.  

Contrastingly, Freya’s blog was effective in creating a blog relevant to Dietitians which was full of relevant and useful information backed by science. These findings portray that AI is not yet capable of writing content creation which is evidence-based and useful to use in practice for healthcare professionals. Thus, healthcare companies should stick to a human writer for evidence-based content that is correctly referenced. 

The results from this study are also supported by previous literature within the scientific writing domain, stating that AI is not yet effective in producing evidence-based content which critiques journal articles or forms an argument. It is, however, useful in producing new ideas and forming a plan for essays which may be useful to those suffering from writer’s block. Furthermore, within the creative writing domain, research suggests AI is already capable of writing poems, chapters in books and blog posts (e.g fashion and news articles). This type of content has been shown to compete against the standards of highly experienced writers.  This is exciting as it showcases AI’s potential in the future for creating content within the scientific domain.  

In summary

Overall, this was an interesting and relevant study to conduct and review as many of our clients could potentially benefit from this tool in the future. As of yet, however, improvements are needed for it to be implemented within the scientific content creation domain and hopefully more studies like this will encourage innovations and progress within the system. 

So we will continue to watch this space, as we continue to learn more about AI’s capabilities and how this will positively or negatively affect the content creators’ human ability to compete.