April 2024
Tinius Digest
Månedlige rapporter om endringer, trender og utviklinger i mediebransjen.
Om Tinius Digest
Tinius Digest gir deg en oversikt over rapporter om og analyser av utvikling i mediebransjen og publiseres en gang i måneden. Her er våre viktigste funn fra denne måneden.
Del gjerne innholdet med kollegaer og bruk det i møter og presentasjoner.
Innhold
- Consumers of alternative media are more diverse than you think
- 7 out of 10 journalists use AI at work
- Urgent need for preparation for AI transformation
- Non-engagement signals can improve content ranking
- Expanding news deserts in Europe
- Public service is becoming increasingly important
- An emerging AI divide
Consumers of alternative media are more diverse than you think
Researchers at Aarhus University have analysed the use of alternative media in Denmark, using data from a survey of 2,455 Danish adults. The study provides the first comprehensive analysis of alternative media consumption in Denmark.
Key findings
1
Demographic profile of users
Alternative media is predominantly used by older men, especially those living in disadvantaged parts of Denmark. However, the appeal cuts across education levels, suggesting that alternative media attract a diverse audience from various social backgrounds.
2
Political interest and ideological orientation
Users of alternative media tend to be highly interested in politics and position themselves ideologically further left or right than the general population. However, they do not necessarily hold extreme political views, and left-wing media users align more with left-wing ideologies, while right-wing media users align with the right.
3
Overlap with mainstream media use
Contrary to concerns about filter bubbles, most alternative media users also engage with mainstream news. This includes a significant number who regularly consume public service and national daily news, indicating that they do not entirely reject traditional media sources.
4
Greater use of right-wing media
Right-wing alternative media are used more widely than their left-wing counterparts. An estimated 16 per cent of the Danish population engages with right-wing alternative media, compared to just three per cent for left-wing media.
7 out of 10 journalists use AI at work
Researchers at the Associated Press, Northwestern University, and the University of Amsterdam have published findings from a survey of 292 media professionals, showing how AI is reshaping news production, ethics, and workflows.
Key findings
1
Widespread use of AI in journalism
73.8 per cent of respondents reported using generative AI in their work, primarily for content production (69.6%), such as creating headlines, social media posts, or drafting articles. Other uses include multimedia creation (20.4%) and business tasks like SEO and marketing (16.6%).
2
Changes in job roles and workflow
Nearly half (49%) of respondents indicated that AI has already reshaped their workflows, introducing new roles such as "AI specialist" and "prompt designer." Tasks now focus more on AI-assisted editing and fact-checking, with efficiency gains still needing further validation.
3
Ethical concerns and supervision
While 81.4 per cent of participants felt knowledgeable about AI, there are growing ethical concerns. The most prominent issues include the need for human oversight (21.8%), accuracy (16.4%), and bias (9.5%). Many fear AI may produce inaccurate or biased content, with respondents advocating for editors and managers to ensure responsible use.
4
Guidelines and transparency gaps
Over 61 per cent of respondents are aware of AI usage guidelines, but many call for more straightforward, more enforceable rules. Some newsrooms have developed internal AI policies, though 14.5 per cent stress the need for adherence to journalistic values such as transparency and accuracy.
Urgent need for preparation for AI transformation
The independent UK think tank The Institute for Public Policy Research (IPPR) has published a report exploring how generative artificial intelligence is expected to transform work in the UK.
Key findings
1
Knowledge work will undergo major transformation
Generative AI is poised to revolutionise knowledge work by enhancing productivity across sectors. AI could increasingly handle tasks like generating text, managing data, and creating strategic content. The report stresses the need to prepare for this shift, as many aspects of routine knowledge work may soon be AI-driven.
2
Back office and cognitive tasks are highly exposed
Jobs involving routine cognitive tasks—especially in back-office roles like administration—are the most vulnerable to AI disruption. About 11 per cent of tasks in the UK economy are already exposed to AI, but this could rise to 59 per cent with further integration of AI into workplace processes. Back-office jobs, often filled by women, face the highest risk of displacement.
3
AI adoption could lead to significant job displacement
In a central scenario, the report suggests that 4.4 million UK jobs could be displaced by AI, particularly in administrative and entry-level roles. However, the economic gains could offset these losses if AI is used to augment rather than replace human work, potentially boosting UK GDP by up to 13 per cent.
4
Women will be disproportionately affected
Women, who are more likely to hold roles in administrative and secretarial jobs, will experience a higher risk of job displacement. About one-third of administrative jobs, primarily filled by women, could be affected by AI, necessitating targeted support for these workers.
Non-engagement signals can improve content ranking
Researchers from a collaboration between UC Berkeley, LinkedIn, and Pinterest have studied the potential of non-engagement signals in ranking content on social media platforms.
Key findings
1
Retention boosted by quality signals
Platforms can increase user retention by including non-engagement signals, such as content quality measures, in ranking algorithms. While engagement-based rankings are effective, incorporating quality proxies (e.g., item surveys and diversity metrics) may lead to better long-term retention.
2
Impact of active user controls
Despite low user uptake, platforms have found that those signals effectively predict content quality and user satisfaction when users actively engage with content controls—such as blocking or hiding content.
3
Generative AI may unlock better ranking systems
Workshop participants discussed the potential of generative AI to improve quality metrics and user control options. Although technical challenges remain, AI could automate content quality assessment and provide users with more intuitive ways to customise their content feed.
4
Transparency and its unclear effect on retention
The study found no strong evidence that transparency or explainability of algorithms directly improves user retention. However, transparency efforts may still contribute to user trust and satisfaction.
5
Diverse engagement as a quality indicator
Content that engages a more diverse audience tends to score higher in quality metrics, suggesting that platforms could benefit from ranking content that appeals to a broader user base.
Expanding news deserts in Europe
This report from the Centre for Media Pluralism and Media Freedom highlights the growing problem of news deserts across Europe—areas lacking sufficient, reliable, and diverse news sources.
Key findings
1
Rural areas most affected by news deserts
In many EU countries, rural regions face significant issues accessing local news. As local media outlets decline, newsrooms become centralised, with fewer journalists on the ground.
2
Economic pressures endanger local media sustainability
The financial situation for local and community media is increasingly unstable. Declining advertising revenues and biased state funding allocation make it difficult for local outlets to survive, particularly in Central and Southeastern Europe.
3
Local journalists face worsening conditions
Journalists working for local outlets face poor working conditions, with many freelancers experiencing job insecurity. There has also been a rise in online harassment targeting local journalists.
4
Inadequate coverage of minority communities
Social inclusiveness is an ongoing challenge for local media, especially in their coverage of minority and marginalised groups. Many outlets fail to engage with or accurately represent these communities.
5
Emerging best practices
Despite these challenges, local and community media succeed through innovative formats such as slow journalism, newsletters, and podcasts. These formats have helped rebuild public trust and engage new audiences.
Public service is becoming increasingly important
The Swedish Independent Public Service Association (Föreningen Oberoende Public Service) has published a report outlining the challenges and opportunities for Sweden's public broadcasting services ahead of a new broadcasting license in 2026.
Key findings
1
Public service essential for democracy
The report argues that Swedish public service media plays a crucial role in fostering "enlightened citizens." The mission to inform, educate, and entertain remains critical for supporting democracy, ensuring access to unbiased news, and upholding freedom of speech.
2
Global media giants threaten diversity
Increasing consolidation of media ownership by American and Chinese corporations risks reducing media diversity. Swedish media, including public service, faces growing pressure from international streaming services and social media platforms that dominate ad revenues and viewer attention.
3
AI's growing impact on media
AI is identified as one of the most significant technological developments of our time, set to transform content production and distribution. Public service media must leverage AI to remain relevant while addressing risks like AI-driven disinformation and manipulation.
4
Securing sustainable funding
The report stresses the importance of maintaining robust, long-term funding for public service, particularly in light of rising operational costs and global competition. The proposed solutions include indexing public service funding to inflation and creating a buffer for unexpected cost increases.
An emerging AI divide
Microsoft Research has published a report mapping the growing disparity in generative AI adoption across different regions and demographics in the U.S. The report analyses search data related to ChatGPT and highlights patterns in awareness and usage of AI tools and how they mirror the long-standing digital divide.
Key findings
1
West Coast leads in AI interest
States like California and Washington show the highest search rates for ChatGPT. At the same time, states in the southeastern (Appalachia) and the Gulf region have persistently low search rates, indicating a geographical divide in AI interest.
2
Education is the strongest predictor
Higher education levels correlate significantly with the interest in AI tools. Counties with more college-educated residents show a stronger tendency to search for ChatGPT, highlighting the role of education in closing the AI divide.
3
Urban areas show more AI engagement
AI interest is higher in urban counties with more technology and creative industry jobs, reflecting socioeconomic divides similar to earlier digital technologies.
4
Racial and ethnic patterns emerge
Counties with higher Asian and Hispanic populations have more searches for ChatGPT, while predominantly White and Black counties show lower engagement with the technology.
5
Digital inequalities are deepening
While generative AI has the potential to be more accessible due to its affordability and multi-lingual capabilities, early data suggests that familiar patterns of digital marginalisation are being reinforced.