Digital Innovation Guest Columns
6 mins read

Democratising AI for all newsrooms: JournalismAI project

When we published the JournalismAI report – New Powers, New Responsibilities – in 2019, we warned of the growing inequalities within the journalism industry. 

A few large organisations may benefit from artificial intelligence and others risk falling behind. Inequalities may also grow between news organisations in more developed economies and emerging ones. We, at JournalismAI, made it our mission to do what we can to mitigate these inequalities by designing targeted initiatives. 

That’s why we launched the AI Academy for Small Newsrooms, with the support of Google News Initiative. The free, 6-week programme aims to support small news organisations that want to learn how artificial intelligence can best aid their journalism. 

The team will select 20 journalists from the Europe, Middle East and Africa (EMEA) region for the programme. The course will be taught by working journalists, researchers and media professionals who are currently leading the adoption of artificial intelligence in the industry. 

The deadline for applications is August 1, 2021. Find more details, including how to apply, on the website

Do you want to exchange ideas, thoughts, and experiences with AI-journalism innovators from across the world? Join the JournalismAI community on Telegram and take part in the conversation.

AI Adoption Challenges

In the design phase of the programme, we spoke to a number of small news organisations across the globe to understand the challenges they face – an independent newsroom in Hong Kong with less than 40 employees; a 3-person team in Mexico; A 10-person newsroom in Lebanon. Besides their relatively small size, these teams have one more thing in common: high AI adoption ambitions. All three are currently taking part in the JournalismAI Collab Challenges. Here’s what they told us:

Initium Media in Hong Kong hopes to develop an AI-based social media listening tool that can analyse online conversations and provide more than just keywords.

DataCrítica in Mexico is already using AI for investigative stories, specifically using NLP and entity extraction.

Maharat News in Lebanon is developing an AI-based social media monitoring software that can work both in English and Arabic.

While these newsrooms recognise the importance of leveraging AI, they also acknowledge that the road to adoption and implementation is not easy for small organisations with limited resources. The issue begins with access to talent and the global tech workforce shortage.

In the UK alone, for instance, about 518,000 additional workers will be needed to fill the highest-skilled digital roles by 2022. As the use of AI across industries gains steam, news organisations are now competing with tech companies, governments, and other industries for a limited pool of talent.

This makes things even more difficult for small news organisations, especially from emerging economies. Gibrán Mena, director of DataCrítica, explains:

“The world is polarised in a way in which the Global North controls technology. Naturally, the place where the first AI teams were born is in the US. But this model of controlling the development of the technology and being able to build teams around it is not something that applies to our reality in Mexico.”

Some news organisations in Latin America try to replicate larger newsrooms’ models of building separate teams of reporters, developers, and data scientists who can work together. But not every organisation has the capacity to do that.

The lack of team members with technical AI skills could deeply impact the AI adoption journey of a news organisation. For example, this could lead to a lack of good data management practices, as explained by Lisa Gibbs, director of news partnerships & AI news lead at the Associated Press: 

“Local newsrooms sometimes don’t have great systems for managing data. They often don’t have the capacity to collect and use data in structured ways to come up with new ideas and insights.”

In organisations that use AI at scale, employees with AI expertise advise on the data acquisition process. They help acquire a greater variety of data sets from different sources. They also have a greater ability to link and analyse the data they collect. This ensures that their companies have good data to work with as they develop new AI solutions. But much of this also rests on the costs involved in recruitment, maintenance, and deployment.

Consider, for example, the use of AI for data analysis and to develop tools for reporters. This use case alone involves very significant costs for a news organisation.

Despite their ambitions, the team at Initium Media remains wary about the cost factor and the return-on-investment. Ning Hui, senior journalist for international news, explains:

“With the current costs, there’s no way I can convince the management that investing in AI will pay off. It’s impossible to allocate investments on this kind of project if short-term results are not guaranteed. If it’s one specific project that guarantees results, we might try it. However, if the solution involved is very costly, even if it might be useful down the line, we may not be able to do it without financial assistance.”

When news organisations are also fighting an existential battle to stay in business, the costs involved in developing and implementing AI make them naturally averse to taking risks and less willing to try new technologies, remarks Gibbs.

Finally, there simply aren’t too many AI adoption playbooks designed specifically for newsrooms. And this lack of guidance for small newsrooms doesn’t make things any easier, according to Layal Bahnam, programme manager at Maharat News:

“Adopting AI is an experimental business. In the end, it’s possible that the solution you build doesn’t meet the challenge it was set out to address. This needs to be acknowledged. The adoption of AI needs to come with a clear methodology so that objectives can be met as accurately as possible.”

Potential Solutions

In spite of all the challenges, there are possible solutions that could help to democratise AI for newsrooms of all sizes – and those often start with collaboration.

Lisa Gibbs encourages newsrooms to take part in international collaborative initiatives with other news organisations, like the JournalismAI Collab Challenges.

Another option is to collaborate with academic institutions that often have more time and resources to develop AI solutions. Early research shows that collaborations like these can increase organisations’ resilience and rewards.

In the US, Gibbs and her team at AP recently received a grant from the Knight Foundation to help local newsrooms overcome barriers to AI adoption and to get them to be “AI-ready”. One thing is clear: despite the many hurdles, giving up on AI is not an option for most news organisations.

Small and local newsrooms already face a number of other challenges in relation to technology – for example with tech platforms’ algorithms deciding which content to show more prominently on a news feed, says Ning Hui. Resisting the AI journey would mean giving up further competitive edge:

“Long-term, if we don’t keep up our unique ways of engaging with audiences, then sooner or later massively-produced content or bigger newsrooms positions will be affirmed, causing smaller outlets to disappear.”

Keeping up with the advancements in artificial intelligence is important for the future of any news organisation, Ning Hui adds. This applies to local brands as well as digital-native organisations that cover niche topics, filling the gaps left by mainstream media:

“To make the landscape sustainable, diverse, and with more voices, smaller newsrooms need to stay at it. Otherwise, it’s not going to be a good future we are looking at for the news industry.”

Lakshmi Sivadas
Community Coordinator, JournalismAI

JournalismAI is a global initiative that aims to create opportunities for journalists and media organizations to come together and explore solutions to improve the future of journalism with AI. It is a project of Polis, the journalism think tank at the London School of Economics and Political Science, and is supported by the Google News Initiative. In the past two years, it published a report on the use of artificial intelligence in journalism, “New Powers, New Responsibilities”, followed by the creation of an online training course to introduce journalists to the potential of machine learning. It is currently conducting the 2021 JournalismAI Collab Challenges.