What screen time really does to children's brains Screen time has become synonymous with bad news - but the science may not be as straightforward as it seems

"Neither Prof Przybylski nor Prof Etchells dispute the grave threat of certain online harms, such as grooming and exposure to explicit or harmful content. But both argue [the] current debate around screen time is in danger of driving it further underground" www.bbc.com/news/article...

Thanks for your work

"The Pfizergate saga ends, for now, with the sad truth that seemingly no one can compel von der Leyen to play by the rules of a democracy, in which laws have to be followed, and officials communications are subject to scrutiny, even when it troubles the powerful" @alexfanta.bsky.social writes.

🧵⬇️

"For the French civil servant, much of the rhetoric around European champions is a convenient distraction for national governments reluctant to reform"

That access to documents is a fundamental right in the EU, but deleting documents without independent oversight carries zero penalties, feels like a massive loophole.

Time to fix this.

There is increasing interest in the possibility of using legislation to proactively shape what we see online. I wrote more here for @politiken.dk about the principled and practical questions - if you pick winners, you also create losers politiken.dk/kultur/medie...

Joint statement by a broad coalition of rightsholders regarding the AI Act implementation measures adopted by the European Commission EPC joins a broad coalition of European and global authors, performers, publishers, producers and other rightsholder organisations, to formally express our dissatisfaction with the published GPAI Code...

Commission pushed ahead despite calls for a pause (including from some big European companies), got key EU tech companies and most of the large US ones to sign the voluntary Code, but judging from this statement they lost much of the creative industries along the way www.epceurope.eu/post/joint-s...

Governments Want to Ease AI Regulation for Innovation, But Do Citizens Agree? | TechPolicy.Press A 2025 survey across six countries finds that citizens are not in favor of deregulation and allowing AI innovation to progress unchecked.

Many governments are talking AI and tech deregulation - but what about the public? New survey suggests citizens across the six countries covered "are not in favor of the recent push for deregulation and allowing AI innovation to progress unchecked". www.techpolicy.press/governments-...

How the public checks information it thinks might be wrong This chapter looks at what people do if and when they want to check something important in the news online that they suspect may be false, misleading, or fake.

Still, many of our respondents do not go to news, and esp. those with lower levels of formal education and those who are more disengaged from conventional forms of politics, are less likely to say they would turn to news media.

Much more here reutersinstitute.politics.ox.ac.uk/digital-news... 3/3

Generally, the most widely identified are news sources that the respondent trusts, official sources, and search engines.

Overall, we find that most people identify two or more sources they might turn to, with trusted news sources generally one of, or the, most widely named. 2/3

Much of the public is concerned about what is real and fake when it comes to online news

So what do people do if & when they want to check something important in the news that they suspect may be false, misleading, or fake?

Wrote about it here reutersinstitute.politics.ox.ac.uk/digital-news... 1/3

Joint statement on Gaza from AFP, AP, BBC News and Reuters

Publisher visibility on Google Search has been falling for years. | Abi Watson Publisher visibility on Google Search has been falling for years. According to SISTRIX data, the Daily Mail is now less than half as visible as it was in 2019—and the Mirror has lost 80% of its visibi...

3) Meanwhile publisher visibility in Google Search continue to decline, Enders Analysis suggests. (Vast majority of the decline predates AI Overviews. So far they seem to mostly impact high-volume, low click-through rate celebrity and informational queries) www.linkedin.com/feed/update/... 4/4

Google earnings keep Silicon Valley’s AI flywheel spinning Capital expenditure on data centres and such trappings this year will now be about $85bn, versus its prior estimate of $75bn

2) ”Google says AI Overviews are driving 10 per cent more queries in searches where they appear and haven’t dented revenue. Paid clicks were up 4 per cent year on year, the company said in a call with analysts” the @financialtimes.com reports on.ft.com/4me7Cww 3/4

Google users are less likely to click on links when an AI summary appears in the results In a March 2025 analysis, Google users who encountered an AI summary were less likely to click on links to other websites than users who did not see one.

1) “Google users who encounter an AI summary are less likely to click on links [than] users who do not see one” and “are more likely to end their browsing session entirely after visiting a search page with an AI summary” @pewresearch.org finds www.pewresearch.org/short-reads/... 2/4

More zero-click searches, same revenue for Google, less visibility for publishers?

Three things to read together on (1) how users respond to AI summaries, (2) what the company says it means for volume of queries and revenue, and (3) outside analysis on publishers’ visibility 1/4

Game on!

BREAKING: The EU Commission has released a mandatory template for AI developers to disclose training data. Unlike the Code of Practice, this is not optional. It could have global fallout, as rights holders abroad might use it to sue over copyright.
digital-strategy.ec.europa.eu/en/library/e...

As always, I recommend reading @ofcom.bsky.social's work. It is careful, serious about engaging with evidence, and, as here, willing to look towards possible futures beyond the status quo. 5/5

Such legislation would be hard in most EU member states, but as and when AVMSD is revised, those of us in the EU too can look forward to discussions of if and how one can intervene politically in prominence on digital platforms - picking winners, but also, implicitly, losers 4/5

There is both a principled decision here about how far one wants politicians involved in platforms’ ranking decisions and pragmatic questions about WHO one would then want to give greater prominence.

We have to keep in mind that attention is likely to be a zero-sum game. 3/5

There are at least 3 different facets to what Ofcom is floating.

1) Boost public service media

But also, in passing, two other intriguing possibilities -

2) Boost new providers with PSM-like qualities made by new non-PSM providers

3) Boost some existing private providers

2/5

What would it look like if politicians used legislation to more actively shape what we see on digital platforms?

@ofcom.bsky.social has suggested UK government “may wish to consider legislation” to ensure greater prominence of specific providers on platforms.
www.ofcom.org.uk/siteassets/r... 1/5

The results are broadly aligned with important work pioneered by a number other academics we build on - part of the added value here from our point of view is the comparative, international design esp. including both Brazil and India to overcome WEIRD bias in the research literature 6/6

The difference between corrections without a link and corrections with a link was not statistically significant. The corrections had no spillover effects on uncorrected true and false posts. 5/6

User corrections had small, and often non-significant, effects on the perceived accuracy of false posts and participants’ willingness to share them. The effects were largest in India and Brazil, while they were smallest in the United Kingdom. 4/6

In the control condition, the posts had no comment. In the correction condition, four false posts included a short user comment debunking them. In the correction with fact-check condition, the user comment included a link to a fact check. 3/6

We conducted a pre-registered online experiment in 2021 on representative samples of the online population in Brazil, India, and the United Kingdom (N participants = 3,000, N observations = 24,000), Participants rated nine social media posts about COVID-19 (3 true, 6 false). 2/6

How effective are user corrections on social media, and does adding a link to a fact check improve effectiveness?

In piece led by @sachaltay.bsky.social we find corrections have small effects, adding a fact-check unlikely to make them more effective misinforeview.hks.harvard.edu/article/the-... 1/6

French and German leaders aim to defuse tensions over fighter jet Disputes over governance and workload plague weapons programme that is crucial to European defence ambitions

“There is no point in making plans that are solely political; one needs a plan that works on an industrial and technical basis.” (Let alone effective demand.)

Useful reading for those thinking about roads towards reduced EU reliance on US and Chinese tech on.ft.com/4kUtnAj