AI in journalism is here to stay. Scandinavian media especially has come a long way – most have already introduced some form of AI. The report Nordic Survey Study of AI in Journalism, published by Medier & Demokrati, charts Scandinavian news media's use of the technology. Here’s what we can learn from their experiences and challenges.
In the report, 51 executive managers from media houses in the Nordic region describe their organisations' relationship with AI and automation. There is confidence and optimism about the opportunities AI can offer journalism among Scandinavian media companies. Notions of AI taking over are rejected, instead the technology is seen as a tool to strengthen and expand human abilities.
40 percent of the media leaders questioned say that their media house lacks strategy in the area and almost 90 percent of them believe that the industry is lagging behind in development. In fact, the report author notes, the media industry is no worse at exploiting the potential of AI than other industries are.
Management has a big impact on the innovation climate – leadership and involvement is crucial to the implementation and success of new data driven solutions. Know-how is most often found in IT, innovation and development departments. Thus, the greatest responsibility for AI projects usually lie with technologists and product owners, whereas the editorial perspective is not always represented to the extent it should be, according to report author and media researcher Jenny Wiik.
The most common uses of AI are for data collection and research. An example are algorithms which automatically search for breaking news online. In Scandinavia, the second most common use is for news production, where robots generate shorter articles and notices based on openly available data such as real estate transactions, sports results or traffic and weather.
Where in the production chain do you use AI today? (percent, several answers possible)
| Idea work | 8 |
| Data collection and research | 66 |
| News production (e g AI-generated articles) | 64 |
| Presentation and design | 25 |
| Distribution | 26 |
| User feedback | 13 |
| None of the above | 13 |
The main reasons for choosing automated solutions are to free journalists from time-consuming routine tasks, but also to increase the production of this type of data driven news. Other important areas include personalisation and curation of the news distribution.
Main reasons for automation (percent, several answers possible)
| Automate routine tasks so that journalists can do qualified tasks | 62 |
| Increased production of news in data driven areas such as finance and real estate | 60 |
| Personalisation of news distribution | 54 |
| Strengthen and develop journalistic methods | 50 |
| Increase reach | 33 |
| Personalisation of ads | 33 |
| Speed up workflows | 29 |
According to survey respondents, the biggest obstacle to AI implementation is that organisations do not have the right skills to develop and implement AI. There is also a lack of knowledge about it in the newsroom more specifically.
In the debate about AI, there's an argument that robots may take jobs and make people redundant, however only 8 percent of respondents, have actually encountered such fear among staff.
Main obstacles to automation (percent, several answers possible)
| Lack the right skills | 55 |
| Lack of knowledge about AI in the newsroom | 51 |
| The management is not sufficiently experienced or driven in the matter | 34 |
| Insufficient strategies | 32 |
| We do not have time to invest in innovation | 28 |
| Lack of financial resources | 26 |
| Resistance to change in the newsroom | 25 |
| Difficulties in collaborating on joint projects | 15 |
| Staff's fear of being made redundant | 8 |
| Difficulty of ensuring ethical standards and publisher responsibilities | 4 |
Today, it is no longer controversial to think of readers as consumers and of journalism as a product. 87 percent of the Scandinavian media leaders answer that they agree that journalism must be based more on user experience and less on "gut feeling". The fact that the user experience has become a more important consideration does not mean that the quality of the journalism necessarily declines or that it becomes less socially relevant – on the contrary. In the survey, a majority state that they think that user data has allowed them to get to know the audience better and to adapt the journalism accordingly. They also believe that the introduction of automation in the newsroom primarily has been focused on journalism, rather than commercial aspect, and it has helped to strengthen the news brand.
Here are a number of statements about AI in journalism, tick the option that best suits your perception. (percent)
| Totally agree | Agrees in part | Partially disagrees | Disagree | |
| The automation we have introduced so far has strengthened our journalistic profile | 47 | 47 | 4 | 2 |
| The news industry lags far behind other industries in the development of AI and automation | 27 | 60 | 14 | |
| Automation is more about the business and less about the journalism | 4 | 31 | 58 | 8 |
| In the future, AI will make many journalists redundant | 10 | 31 | 59 | |
| AI is just a tool among others, humans will always be in control | 56 | 33 | 11 | |
| Algorithms make journalism more relevant to users | 12 | 54 | 28 | 6 |
| Automated personalisation makes journalism more relevant to users | 42 | 48 | 8 | 2 |
| AI improves the journalistic quality | 37 | 55 | 6 | 2 |
Part of the survey covers collaboration, both between different parts of the company and between different competencies. While there is a willingness to work together, 60 percent of respondents think that it is still difficult to get technicians and journalists to speak the same language. At the same time, it is clear that there is potential for improvement. 74 percent believe that the organisation needs to learn methods for collaboration.
Development of AI usually requires collaboration between different roles and competencies, how do you think it has worked in your organisation? (percent)
| Yes | No | Uncertain | |
| It is difficult to get technologists and journalists to speak the same language | 60 | 32 | 8 |
| We have received (internal or external) help in facilitating the collaboration | 43 | 42 | 16 |
| We need to learn methods for cooperation | 74 | 12 | 14 |
| The collaboration between different competencies has strengthened our journalistic brand | 76 | 8 | 16 |
| We have a good working climate where it is ok to make mistakes | 78 | 8 | 14 |
| Collaboration between different parts of the company can threaten the editorial autonomy | 10 | 80 | 10 |
| We have clear roles and everyone knows what to do | 41 | 39 | 20 |
| We miss new ideas due to poor internal communication | 28 | 47 | 26 |
| A large part of the development work is outsourced | 20 | 76 | 4 |
AI can mean huge benefits for media companies, both competitively and from a publicist's perspective. It is not only about technology but also about the organisation's ability to benefit from technology. Therefore, AI is not just a matter for the technology department, but concerns the management and the editorial managers to a very high degree. However, over 40 percent of media leaders state that their organisation does not have a specific strategy for developing and implementing AI.
It is mainly strategy and innovation managers and product owners who take responsibility for initiating and driving innovation. The journalistic leadership generally plays a smaller role, which means they may not impact the internal innovation processes to the desired extent.
Who has the main responsibility for developing and implementing AI in your organisation?
| Strategy / innovation manager or similar | 55 |
| Product manager | 49 |
| Editor-in-chief / publisher | 26 |
| IT manager | 26 |
| News editor | 15 |
| Internal project manager | 13 |
| Other | 11 |
| Online / web editor | 8 |
| External consultant | 6 |