Desember 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
- The risks of online harm to children are increasing
- AI disclaimers on political ads may backfire
- Video games enhance cognitive performance
- 6 in 10 adolescents think there is too much advertising on social media
- Social media can predict mental health crises
- AI chatbots reduce public knowledge sharing on online platforms
- Video games provide a (short-term) boost in mood
The risks of online harm to children are increasing
The Norwegian Consumer Council (Forbrukerrådet) has published a report about how children and adolescents are exploited online and offers recommendations for creating a safer digital environment.
Key findings
1
Commercial exploitation
Social media platforms exploit children’s vulnerabilities with surveillance-based advertising, using personal data to target ads that reinforce insecurities, such as content promoting weight loss or cosmetic surgery. These practices often blur the line between entertainment and commercial messages, making it harder for children to differentiate between them.
2
Amplification of toxic content by algorithms
Algorithmic recommender systems optimise engagement by amplifying extreme or disturbing content. This “rabbit hole” effect disproportionately affects adolescents, exposing them to self-harm content or radicalisation material. Enhanced user control and changes to recommender systems are recommended to reduce these harms.
3
Addictive design practices
Many platforms employ features like infinite scroll, autoplay, and “streaks” to maximise engagement. These mechanisms exploit psychological vulnerabilities, particularly in young users, leading to excessive screen time that can negatively impact mental health, social skills, and education.
4
Limited efficacy of technical barriers like age verification
Methods such as ID-based age verification and AI-driven age estimation often fail to protect children adequately and come with risks such as privacy violations and digital exclusion. They may also push children to less moderated platforms, exacerbating risks.
AI disclaimers on political ads may backfire
The Center on Technology Policy at NYU has studied the effectiveness of state-mandated AI disclaimers on political ads.
Key findings
1
AI disclaimers reduce trust
Political candidates whose ads contained AI disclaimers were rated as less trustworthy and appealing. Respondents also found these ads less accurate and were more likely to report them on social media.
2
Weaken the impact of attack ads
Labels not only lowered trust in the candidate who created an attack ad but also did not affect the candidate being attacked. Instead of increasing scepticism about the ad’s claims, the disclaimers resulted in a “backfire effect,” making the attack ad’s creator look worse.
3
Political bias affects how disclaimers are perceived
AI disclaimers had the strongest impact when ads came from candidates of the viewer’s own party or had no clear political affiliation. However, disclaimers had little effect on candidates from opposing parties, as respondents already viewed them unfavourably.
4
Many viewers do not notice disclaimers
Up to 37 per cent of respondents did not recall seeing an AI label, suggesting that disclaimers may not be as effective as intended. Their impact could be even smaller in real-world scenarios where attention to ads is even lower.
Video games enhance cognitive performance
Researchers from Western University in Canada have conducted a global study on over 1,000 participants to examine the effects of exercise and video gaming on cognitive and mental health.
Key findings
1
Exercise improves mental health but not cognition
People who engaged in regular physical activity showed better mental health outcomes, with lower levels of depression and anxiety. However, there was no significant link between exercise and cognitive performance.
2
Video games enhance cognitive performance
Participants who played video games regularly performed better on cognitive tests, particularly reasoning, memory, and processing speed. However, gaming was not associated with mental health improvements.
3
Different lifestyle factors support different aspects of brain health
The study suggests that individuals can optimise their well-being by tailoring lifestyle choices: exercise for mental health benefits and gaming for cognitive gains. These findings challenge common assumptions that both activities improve all aspects of brain health.
4
Gaming frequency matters
Frequent gamers (playing more than three hours per week) outperformed infrequent gamers and non-gamers on cognitive tasks. This suggests that the cognitive benefits of gaming increase with more regular play.
6 in 10 adolescents think there is too much advertising on social media
The Norwegian Media Authority has published the results of a survey examining how Norwegian children and teenagers (aged 9–18) engage with online advertising and influencers.
Read more (in Norwegian).
Key findings
1
Children feel overwhelmed by online advertising
64 per cent of young social media users believe there is too much advertising on digital platforms. This refers to content the children perceive as advertising.
2
Strong influence on young audiences
48 per cent of 9–18-year-olds regularly watch content from influencers on YouTube, TikTok, or other platforms, with humour being the most popular genre. A quarter of them follow gaming, music, sports, or fitness content.
3
Children are sending money to influencers
Seven per cent of children and teenagers have sent money to influencers. The main reason (48%) is that they like the creator and their content, while a quarter (24%) do so in exchange for shoutouts. Boys are more likely than girls to send money, and the likelihood increases with age.
4
Heavily exposed to gambling and weight-loss ads
Among 14–18-year-olds, 60 per cent have encountered online gambling ads, while almost half have seen ads for weight-loss products and muscle-enhancing supplements. Even among 9–12-year-olds, 20 per cent report exposure to weight-loss ads, and 25 per cent have seen muscle-building product promotions.
Social media can predict mental health crises
Researchers from the Institute of High Performance Computing and the Ministry of Health in Singapore have studied how social media indicators can help predict public mental health needs during crises.
Key findings
1
Twitter emotions predict mental health needs
Joy, anger, and sadness expressed on Twitter could predict changes in mental health-related emergency visits and online crisis cases. Joy intensity was a strong predictor of psychiatric emergency visits, while sadness levels on Twitter correlated with crisis assessments on the mental health portal.
2
Social media indicators outperform traditional pandemic data
Emotions from social media were better predictors of mental health needs than traditional COVID-19 indicators like daily cases and deaths. Government announcements had some predictive value, but most situation indicators failed to anticipate mental health crises.
3
Twitter activity helps improve crisis forecasting
A forecasting model that included Twitter activity, particularly tweet volume and emotion counts, performed better in predicting emergency room visits and online crisis cases than models relying only on past mental health data.
4
Public trust in social media sentiment for mental health monitoring
The study highlights how real-time social media data can be used as an early warning system for public mental health needs. By tracking emotions online, policymakers can allocate resources more efficiently and respond proactively to mental health crises.
AI chatbots reduce public knowledge sharing on online platforms
Researchers from University College London, LMU Munich, Corvinus University of Budapest, and other institutions have examined how the introduction of ChatGPT has impacted activity on online Q&A platforms, particularly Stack Overflow.
Key findings
1
25 per cent decline in activity
Within six months of ChatGPT’s release, posting activity on Stack Overflow dropped by 25 per cent compared to similar platforms less affected by ChatGPT. The decline was particularly strong for widely used programming languages like Python and JavaScript.
2
No major impact on post quality
While fewer questions and answers were posted, there was no significant change in post quality, as measured by user voting. This suggests that ChatGPT displaces posts across all quality levels rather than filtering out low-quality content.
3
Experienced users posting less
The decline in Stack Overflow activity is not just due to beginners finding answers elsewhere. The study found a significant reduction in contributions from both inexperienced and expert users, raising concerns about the sustainability of public knowledge-sharing platforms.
4
Potential risk for AI training data
Large language models rely on publicly available data for training. The reduced contribution to Stack Overflow and similar platforms could limit the availability of high-quality training data for future AI models, creating a paradox where AI tools erode the very data they depend on.
5
Shift towards private knowledge silos
As users turn to ChatGPT instead of open forums, knowledge-sharing becomes more privatised. Unlike public Q&A sites, AI-generated responses are not publicly archived, reducing the collective pool of freely available knowledge.
Video games provide a (short-term) boost in mood
Researchers from Tilburg University, the University of Oxford, and the Karolinska Institute have examined how playing video games affects players’ moods.
Key findings
1
Video games can slightly improve mood
On average, players reported a 0.034-point increase in mood (on a 0–1 scale) while playing. The study estimated that around 72.1 per cent of players experienced this uplift in mood.
2
Most improvement within the first 15 minutes
The biggest increase in mood occurred within the first 15 minutes of gameplay. After this, mood levels stabilised but did not return to pre-play levels. Players who started their gaming session in a lower mood improved more than those already in a good mood.
3
Comparable to other leisure activities
The mood improvement from playing was similar in scale to activities like watching TV or reading. However, it was smaller than activities like listening to music, walking, or exercising. The study did not determine whether this mood boost lasts beyond the gaming session or accumulates into long-term well-being benefits.