Researching the impact of Artificial Intelligence on publishing

Catherine Etienne, Senior Associate at Frontier Economics, writes about her experience researching and writing the People Plus Machines: The Role of Artificial Intelligence in Publishing report:

Our research on the impact of Artificial Intelligence (AI) on publishing – conducted prior to the Covid outbreak – shows that these are exciting times for the industry. But what do we mean by AI? Our discussions with publishers highlighted that there is no standard definition of AI being used in the sector right now. Broadly, it covers a wide range of technologies, including machine and deep learning, which are often combined with text or image recognition technologies to enable computers to understand, learn from and act according to the information they are given.

Our industry survey and consultation process suggest that, despite AI investment activities gaining momentum only fairly recently, since 2017 or so, we can expect to see a wide variety of AI technologies being applied by UK publishers at every stage in the value chain within the next two years.  Some examples from our research include:

• AI tools for authors to identify, compile and summarise research materials;
• solutions that enable publishers to carry out accurate and fast language quality and plagiarism checks on their content;
• generating insights to improve marketing strategies and improving the quality of service publishers provide to customers and suppliers.

By investing in this way publishers are generating significant benefits: two-thirds of the large, currently investing publishers we surveyed told us they are already realising benefits. There will also be wider benefits for customers, including academic  researchers, authors, students and teachers. 

As our case studies demonstrate, AI has the capacity to perform routine repetitive research and summarisation tasks faster and more accurately than humans. For publishing, this could help make researchers and teachers more productive, freeing them up to spend more time analysing, creating or  teaching.  This suggests that by increasing their investment in AI, publishers should be able to generate increasing benefits to society.

Unsurprisingly, investing in AI is not without its challenges and publishers reported facing a wide range of interesting issues.  For larger publishers, the key challenge appears to be technology focused.  Two-thirds of large publishers said that acquiring the necessary AI technical skills was a significant barrier to investment.  This may explain why we observe significant collaborative activity between publishers and AI research organisations, such as universities and private AI-tech companies.  Another key challenge relates to integrating new AI solutions with existing IT infrastructures, and getting buy-in to investment decisions given the fast pace of change in AI technologies. For smaller publishers, the challenge is overcoming the significant upfront costs of investing in AI, including knowing where and how to invest. 

Identifying these challenges offers an opportunity to help the publishing industry overcome investment blocks and generate further benefits to society.  Our report sets out a number of recommendations on where to focus over the next couple of years, including providing greater investment certainty through a stable intellectual property framework, promoting continued R&D collaboration between publishers, AI-tech SMEs and academia, and helping smaller publishers access the finance and skills they need to start investing in AI.

For more thoughts from Catherine on the People Plus Machines: The Role of Artificial Intelligence in Publishing report, view this video: