What is the future of news after Covid-19? How will new technologies such as AI change our journalism? How will forces for social change such as BLM impact on that process?
Polis director Charlie Beckett argues that these are all related trends in this article based on the evidence he gave recently to a UK parliamentary inquiry into the future of journalism.
Last year, Polis conducted a global survey of what newsrooms around the world were doing with AI and machine learning technologies and we are now working with a network of 1000+ digital news media people on creating training and innovation projects. So we are in a great position to see how Covid-19 is impacting on technology-driven innovation. The journalists we talk to are even more enthusiastic about finding out about AI because they can see that it can help cope with the business and editorial challenges they face.
Covid-19 is a human story about health, economics and social justice but it’s also a massive data story. AI helps to gather, analyse and present that data in useful, relevant ways that audiences can connect to more easily. It also offers much greater efficiency and effectiveness for newsrooms as a whole and support for individual journalists to do their research and content creation.
But, of course, journalists also tell us that resources are being hit hard, which limits the ability for individual news organisations to develop new tools or create new systems. AI can be complex and you need specialist skills to train data or to use the algorithms well. A lot of projects are being put on hold or even cancelled. That’s why news organisations seem so keen to get involved in collaborative work so they can learn best practice from other experts or practitioners. The post-Covid-19 media world – like the rest of society – is going to be more data-driven, more algorithmically-powered. So it’s vital for journalists to get across these trends. But they can’t do it alone anymore. They need support to protect journalism values and to innovate.
We need to put this immediate crisis in the historical context of what has happened to the news industry in the last couple of decades and to think what it means for the future of journalism in the next period.
Changing Information Ecosystem
Journalism has been through an extraordinary, multifaceted series of changes and they are not all about technology. The whole information ecosystem has changed and so has society. In terms of technology and news, there have been three big shifts. First, there was the move online more than 20 years ago. News organisations got websites. Secondly, there was the move into social media. This was much more complicated in the ramifications and in terms of the kind of journalism that was produced and the relationship with the audience. Thirdly, we are now in a phase where journalism is going to become much more distributed, much more diverse in terms of the organisations, and much more devolved in terms of how people get their information. AI and ML technology is going to be at the heart of this.
It is important that we move this debate about supporting journalism away from the old idea of trade-off between, for example, the technology companies and the news industry. That really misunderstands where the success and innovation of journalism and its problems come from: they are structural. I could give examples of clever new tools for specific uses, but it is even more important to see how the better news organisations are using the technologies to redefine their business model or their relationship with audience and the way they create value. That then tells us how we, as a society, might support that process.
Our report on AI and journalism has lots of case studies about how newsrooms are using the technology, but we tried to stress the structural shifts that are happening too. So if you want to help journalists to do data journalism, or to use audience data to improve subscriptions, then the best way you could help them would be to support newsrooms overall capacity. Improve their technology knowledge and training and facilitate news media’s collaboration with the universities, start-ups, tech companies or even each other.
Structural Support for News Media
That is what we are doing with JournalismAI. We are trying to get newsrooms to collaborate and share best practice with the network. We have created an online training module to help them create an AI strategy. We do not have the money to actually make everything happen. But we hope to grow our global network and spread ideas and innovation organically.
The key thing about AI is that it is a pervasive technology. Look at data, for example. Data can mean brilliant investigative journalism, like the Panama Papers. It can mean extensive data sets on homelessness that local media might be able to dig into. Or it might mean data about your audience, so that you can give them content that is more relevant to them and that they will value and therefore pay for.
Journalists tell us that while AI does replace some tasks and even create new tools, products or services, it is mainly augmenting what they would like to do. It is making them more effective and efficient. The biggest problem they face is creating a strategy around that, having the resources and the knowledge from the management right through to frontline reporters. Newsrooms need specialist knowledge to understand the algorithms and the black boxes. But they also need to work out the capabilities and limits of some of this tech. They need to work out how it fits best with their editorial values and practice.
Inequalities and Public Information
We have also identified serious inequalities around AI. There are inequalities within the news industry. Some large companies are much more advanced and have greater R&D capacity. The news industry as a whole is tiny compared with other industries and sectors that might use AI: pharmaceutical, retail, health. To get the technologists to pay attention, the news media has to do a much better job of explaining its needs and the special value and opportunities journalism offers. It is not just about convincing the tech companies to hand over money, because the tech companies are actually being quite helpful, even if you suspect their motives to be self-interested. The news media needs to leave taxing the tech companies to governments and focus on how the platforms and network companies might provide structural help such as sharing data and promoting quality journalism.
— Declaration of interest: JournalismAI is funded by the Google News Initiative.—
Of course I would love it if there was also more public support for universities, institutes, NGOs, and even start-ups when they try to provide the infrastructure for journalism to exploit those technologies. It should not be seen just as a subsidy for a particular news organisation. It should be seen as an investment in public information.
The optimistic scenario—and the evidence is that it has happened already in the last decade or so—is that AI can take at least some of the automated labour away and leave the journalists to do the things they are best at, which should include curiosity, passion, human interest, but also specialist knowledge or new formats. That is the new skillset. It’s not just about understanding the algorithms. There are new roles and responsibilities. New job titles are emerging, such as “algorithm editor”, which is someone using human judgment to oversee how well the machines are working.
It’s what AI-journalism scholar Nick Diakopoulos calls the ‘hybrid journalist’. If your job can be replaced by an algorithm, you have to question what you were doing in the first place to add value as a human being. Now ask if the technology can enable you to do something better.
One of the biggest skills journalists are going to need is to be entrepreneurial—I mean that in the broadest sense—to understand the diversity of journalism. There is an over-abundance of information in the world at the moment, and too much of it is the same or of poor quality. One of the things that journalists are going to have to do is work out how they add value in what they do, day to day and year by year. That has only partly to do with the technology, of course. You get pioneer journalists such as Sophia Smith Galer at the BBC, who is doing clever work on TikTok. I do not think TikTok is the solution to the news media’s problems, but she is a good example of a digital-savvy person using technology, and she has done it in an entrepreneurial, creative way in line with the core values of her employer.
Is there a big skills gap? Yes, definitely. The biggest conclusion in our JournalismAI report was that there is a huge skills gap, not just in that specialist tech knowledge, but in understanding the new systems: for example, what audience data might tell you about how you relate to the public and what you might do with that. It’s just a small step, but we have created another online training course, which is partly about strategy and is going to move on to be more nuts and bolts. We designed it with journalists for journalists, because newsrooms told us that basic knowledge is needed before you can introduce the technology. As an industry, we are not good at thinking in that strategic, systematic way about the resources that our journalists need.
From an industry point of view, the skills gap is also related to the lack of diversity. Journalism needs people with the skill of understanding different communities, identities, or classes. It needs people with different life experiences and perspectives. We also need more people who understand science or law, for example. In this recent crisis we have seen the deficit of understanding when it comes to statistics, health science and epidemiology. In that sense, I would say, if you want to become a really successful journalist, study law, languages or science.
Lawyers, for example, do continuous career enhancement or professional development. Journalists often have fantastic experience and have developed great contacts and relationships, and yet they need to be retooled or refreshed. That is not just for the technology but for the new circumstances they are operating in.
The Human Factor
Journalists need core human aptitudes such as curiosity, creativity and commitment. You have to believe in the ideal that your journalism is going to have some kind of impact. To reflect back on society in its diversity, you have to have lots of different approaches, attitudes and skills. The idea of the hard-boiled hack is one useful myth. I am not saying it is wrong, but especially in the modern setting, where journalism cannot just rely on being that cliché core industry, it has to be more diverse in its make-up and its practice.
Journalists can be helped by technology but they need to improve their human values, their emotional literacy. By that I mean understanding identity and practicing empathy. Not as a nice thing to have but as a central organising principle. If you are going to try to connect to people, you have to be able to understand their lives. Understanding might mean at a human-interest level. It might mean understanding at an expert, curational level. Understanding what people do with the news. There is so much room in this industry for different kinds of skills and competencies. The problem at the moment is the reverse. It has become demographically more homogenous.
The pandemic and related social change, such as the current protests around race, are having immediate effects but they will also change our lives in the long-term. The negative impact of the economic disaster it has caused will hit the news industry hard. But it is also going to accelerate some existing trends, for example, around working practices. Combined with the anger over issues such as racial injustice and it is clear that journalism will need to adapt its culture as well as its business models. We don’t know exactly how, but news media will certainly have to understand and reflect those changes. Technology is only part of the story of the future of news, but it will continue to be a critical catalyst for structural change.
This article is by Professor Charlie Beckett, Dept of Media and Communications, LSE. He is director of Polis, the LSE’s journalism think-tank and leader of the Polis Journalism AI projected, supported by the Google News Initiative. This article is based on oral evidence Prof Beckett gave to the House of Lords Communications and Digital Select Committee inquiry into the future of the news industry. You can read a full transcript of Professor Beckett’s evidence here.
The views represent those of the author, not Polis or the LSE