top of page
Zdjęcie autoraMichał Samojlik

How newsrooms are leveraging AI





We already know quite a lot about AI and its capabilities and their impact on editorial operations. Yet, the discussion must move beyond mere awareness to concrete initiatives preparing newsrooms for an uncertain AI-mediated future.


It has been captivating to watch the media experiment with ChatGPT's writing capabilities since its release in November 2022 (and its updated version in March 2023).


We witnessed the world’s first magazine written and designed by AI, the first AI-generated photo to win a world award, the first AI-driven features in newsrooms’ CMS systems, and the first AI models fed with specific data.


Many experts saw the release of ChatGPT and similar language models as an accumulation of work over the last few years rather than a sudden miracle. We realised that we have been using AI for years – and that the core of the current revolution is that AI is now affordable and easily accessible to almost everyone.


In the first half of 2023, journalism grappled with questions like “Will it outperform humans?” “Will it replace journalists?” (some media owners said it would) "What can it really do?”.


However, concrete initiatives to implement AI in newsrooms were rare.


Today, we know much more about generative AI capabilities and their impact on newsroom operations.


Technological challenges remain


According to the LSE's just-released global journalism AI initiative, “Generating Change”, conducted between April and June, 85% of newsrooms have experimented with generative AI technology to varying degrees to date. Yet four in ten news organisations have not significantly changed their approach to AI in the newsroom since 2019.


Newsrooms struggle to allocate the necessary resources and lack the technical skills for successful AI integration. Cultural and ethical challenges remain noticeable hurdles, and there is an urgent need for guidelines, standards and regulations to ensure the fair use of AI in newsrooms.


A screen from LSE global survey of what news organisations are doing with artificial intelligence – the most pressing challenge for newsrooms is of a technical nature.


In his latest article, David Caswell has done a remarkable analysis of media strategies for AI implementation, from efficiency-oriented to product strategies. He notes that the practical use of AI in newsrooms must be supported by infrastructural and organisational requirements. “Any strategy for adapting a newsroom to an information ecosystem driven by generative AI is of little use without concrete, practical projects that translate that strategy into useful outcomes”, David writes.


Here’s the catch. To successfully implement AI tools into daily workflows, every newsroom needs an interface between software and journalists – such as AI features or AI wizards within the CMS (or even an entirely new form of CMS designed specifically for AI workflows). A good example of this is a POC we developed at Autentika, which is presented in the following video:





Personally, I don’t see AI as a shortcut to success. On the contrary, the road seems long and winding in many places. For many editors, AI is tempting, but board members try to be honest with themselves. They know that even after a superficial analysis of the state of the tools available in the newsrooms, they should not dream of a big change. At least not right away.


Innovation managers in media companies want to make room for automation, the use of analytics and generation engines, and integration with an ever-widening range of media products. To do that, they need to take a critical look at the existing publishing software: UX architecture, stability, technical equipment, team skills.

Many companies have accepted the technological limitations of their software. They are aware that their tools are not very efficient. At the same time, they cannot be indifferent to new possibilities. This is a critical moment when they either pay off the debt or accept the competitive advantage.


This is not just my view; we mentioned these challenges in April in our interview with Charlie Beckett about 7 bitter pills any newsroom that wants to use AI must swallow, and many media practitioners repeatedly pointed them out.


Key areas for the use of AI in newsrooms


Nevertheless, many newsrooms – smaller and larger – are already using AI and have integrated it into their daily operations.


Three main areas where AI is being used are:


  1. News gathering – using tools that enable optical character recognition (OCR), speech-to-text conversion and text extraction. AI automates transcription, extracts text from images and structures data after capture. In addition, newsrooms use AI applications that can sift through large amounts of data and recognise patterns.

  2. News production – most journalists now use some form of AI in content creation, with functions such as summarising, paraphrasing, fact-checking, proofreading and SEO tools.

  3. News distribution – here journalists mention AI-powered tools for social media distribution or features that enable greater personalisation.

Respondents to the LSE survey expect AI to influence four main areas:


  • Fact-checking and disinformation analysis

  • Personalisation and automation of content

  • Text summarisation and generation

  • Use of chatbots to conduct pre-interviews and gauge public opinion on specific issues


Real-life examples in “The Newsroom of Tomorrow”.


I realise that the topic is vast and a short article is not enough to cover it. This is also true for the entire section of “Newsroom of Tomorrow”. But we would like to show you very briefly some real-life examples that illustrate the successful use of AI in three key areas:



I hope you enjoy reading – and if you want to exchange ideas, contact me on LinkedIn.





Comments


bottom of page