Fenimore Harper—written evidence (FON0016)

House of Lords Communications and Digital Select Committee inquiry:

The future of news: impartiality, trust and technology

Fenimore Harper communications

1. Fenimore Harper communications is a digital media consultancy offering strategic advice, bespoke media monitoring and training in digital media and generative artificial intelligence.
    
2. This year Fenimore Harper has published reports on scams and disinformation powered by generative artificial intelligence.
    
3. Marcus Beard is the founder of Fenimore Harper communications. Prior to founding the company Marcus worked in the Cabinet Office heading up their digital media monitoring unit. He specifically worked on monitoring harmful health narratives during the COVID-19 pandemic.
    
4. This submission seeks to provide Fenimore Harper’s view on a small subset of the questions, sharing their expertise and experience on digital media.

What impacts (positive and negative) do large technology platforms and online news aggregators have on the UK’s news environment, including media plurality? And how might this change?
 

5. Social media algorithms prioritise engagement and can amplify false or misleading information, making it difficult for the public to differentiate between credible and non-credible sources.

6. Emotionally charged messages on social platforms are shared more widely than neutral ones.[1], [2] Since these platforms began to send a large proportion of traffic to news websites, an incentive has been created to optimise for these emotional qualities.

How is generative AI affecting news media business models and how might this evolve?
 

7. Many ‘made-for-advertising’ news sites have been created, facilitated by Large Language Models such as OpenAI’s Chat GPT-4. These sites have no human oversight. They are entirely automated: firstly, traditional news sites are scraped using RSS feeds of traditional web crawling; article text is then fed to ChatGPT-4 and prompted to be rewritten; these re-written, ‘laundered’ articles are uploaded, published and promoted on social media en-masse. Due to this process, these are low-quality and often inaccurate.
    
8. ‘Made-for-advertising’ content also takes the form of video and audio content made possible by image technology such as stable diffusion, and ‘deep-fake’ audio and video platforms. Recently a large amount of science content on YouTube was found to be factually inaccurate.[3]

February 2024

2