New Powers, New Responsibilities: A Global Survey of Journalism and Artificial Intelligence

"The world that journalism inhabits, especially the media and technology fields, will be reshaped by AI. If journalism wants to stay at the cutting edge, it needs to inform itself about AI and explore its uses and consequences for the newsroom as well as wider society."

This report provides an introduction to and discussion of journalism and artificial intelligence (AI), which may be defined as a collection of ideas, technologies, and techniques that relate to a computer system's capacity to perform tasks normally requiring human intelligence. Funded by the Google News Initiative and carried out by the international journalism think-tank, Polis, the report is based on a survey of 71 news organisations in 32 different countries regarding AI and associated technologies. A wide range of journalists working with AI answered questions about their understanding of AI, how it was used in their newsrooms, and their views on the wider potential and risks for the news industry.

Following an executive summary and introduction, chapter 1 deals with what is happening now: How are newsrooms using AI, and what are their motives for using it? What works and what doesn't, so far? Chapter 2 looks at strategy: What kind of approaches to AI are taken by news organisations? What impact can AI have, and what are the key challenges? Chapter 3 explores ethics and editorial policy: What positive and negative effects does AI have on the quality of journalism? What issues are raised around bias or accountability? What is the role of the technology companies? Chapter 4 examines the role of AI in the future of journalism: What uses of AI do newsrooms aspire to? What kind of skills, roles, training, and education do we need in the future? What is the role of collaboration? How will AI shape the way that journalism is going to evolve as a practice and a business? A conclusion, glossary, endnotes (with links to case studies), and a list of suggested readings and resources complete the report.

As the report outlines, the technologies that come under the umbrella term of AI range from everyday functions such as search, to complex algorithms drawing upon deep learning to create text or videos. The newsrooms surveyed said they are using AI for:

• News gathering (just under 50%): sourcing of information, story idea generation, identifying trends, investigations, event or issue monitoring, extracting information or content.
• News production (about two-thirds of respondents): content creation; editing; packaging for different formats and platforms; text, image, and video creation; repurposing content for different audiences.
• News distribution (just over 50%): personalisation, marketing, finding audiences, understanding user behaviour, monetisation/subscriptions.

"Under-pinning all AI processes is data: data about audiences such as their behaviour; data about the reported subject such as official records; data about the journalism such as sentiment and language. Whatever the precise category, it is clear that the use of AI is permeating newsflows. That makes it harder to identify and evaluate what it does, but it also reflects the adaptability of the technology and its increasing integration into operations."

Survey respondents observed that AI technologies are rapidly developing, alongside other changes in media production and business models. AI it has the potential for wide-ranging and profound influence on how journalism is made and consumed, according to the report. When fully integrated, pervasive, and operating at scale, AI could have value in areas such as audience engagement, story discovery, and business efficiency. Some of the possible uses, such as automated translation and text generation, may enable advances into new areas of journalism, marketing, and product development. Furthermore, AI can free up journalists to work on creating better journalism at a time when the news industry is fighting for economic sustainability and for public trust and relevance. It can also help the public cope with news overload and misinformation and to connect them in a convenient way to credible content that is relevant, useful, and stimulating.

However:

• Even the newsrooms the Polis team surveyed that are furthest ahead in the adoption of AI described it as additional, supplementary, and catalytic, not yet transformational. This is partly because of the special nature and needs of journalism but also because of the relative lack of resources for research and development (R&D).
• Challenges to adoption include: (i) costs, (ii) knowledge/skills, and (iii) newsroom culture/fear of job displacement. AI is a fast-moving field where late adopters could face a hard time catching up, the study cautions. Scale is an issue, too: "...journalism is a relatively small business and AI is a big and expensive technology. News organizations will once more have to show imagination as well as determination if the new AI-powered augmented journalism is to thrive."
• Specific areas of concern: (i) algorithmic bias, (ii) deep-fakes/misinformation, (iii) transparency, (iv) employing AI tools purely for financial gain, and (v) tech companies as potential competitors or R&D gatekeepers.
• AI brings with it editorial and ethical responsibilities (see chapter 3). For example, there were fears that AI might accentuate inequalities between news organisations, with implications for media diversity and the sustainability of "quality" journalism production. To cite another: Algorithmic personalisation could further entrench what cognitive psychologists call "confirmation bias", the tendency according to which people prefer content that reflects rather than challenges their own beliefs and values.

Only 37% of respondent companies had an AI strategy at the time of the survey. Key elements that emerged from the research to help in preparing one include:

1. Assess your stage and state of AI readiness.
2. Understand and categorise the kind of AI technologies you are considering.
3. Decide how AI might relate to your brand and general strategy, the problems it might solve, or the needs it could meet.
4. Evaluate what areas of your organisation might use AI and why.
5. Identify key obstacles: resources, skills, culture, management, etc., and plan how to address them in a systematic way.
6. Assign roles and responsibilities and create a communications structure across the organisation to include all stakeholders.
7. Establish systems of monitoring and reviewing performance and priorities.
8. Create a role for external relations with partners, clients, and wider AI resources with a mission to investigate and incorporate AI innovation.

The biggest future wish from respondents was for general education and specific training for AI in the areas of: (i) AI literacy, (ii) AI skills (basic and advanced), (iii) management awareness, (iv) ethics, and (v) general AI insights. The 3 most common areas for respondents' future AI-tool wish list were for: (i) more automatic tagging/entity extraction (news gathering); (ii) better machine-generated content (news production); and (iii) better personalisation/recommendation engines (news distribution). There was a strong interest in AI that can help with the relationships between the newsroom and content, or newsrooms and audiences, in more nuanced ways.

The report stresses the value of collaboration to make progress in a timely way, with: other news companies; technology companies, academia, and civil society organisations; or industries in other sectors. This means organisations will need to get past typical "ultra-competitiveness" and insularity. One organisation set up specifically to collaborate is the International Consortium of Investigative Journalists (ICIJ), which acts as a resource for training and collaboration more generally and is an example of an intermediate body funded independently able to provide resources and expertise on AI-related journalism.

A concluding thought: "Newsrooms need to make serious investments to attract the skills, knowledge, and innovation that journalism needs to optimise the opportunities of AI and reduce the potential harms."

Editor's notes:

• You may also be interested in "7 things to consider before adopting AI in your news organisation: A training module for newsrooms" (December 12 2019), which complements the report.
• Click here to access a bibliography on journalism and AI (continually updated)

.