Hello you,
Happy new year. I hope it finds you well, energised for the year ahead, and not too hungover.
I’ve written a “reflections on the year” piece ever since I left the UK government in 2021. In that time I’ve reflected on: going into, and then leaving, freelance work; starting, and then ending, a business; and moving to Berlin. I’ve found it a useful exercise, and so here’s another for 2024. Previous ones are here: 2023, 2022, 2021.
2024 was for me – in contrast to the world at large – the first “normal” (ish) year in a long time. There was no global pandemic, I had secure immigration status (see here for my piece on not having that), did not move address 8 times in one year, and stayed with the same employer (AlgorithmWatch, who I joined in September 2023). The biggest change for me, I would say, was AlgorithmWatch promoted me to their Head of Tech Research. But even that had strong elements of continuity, though also meant thinking more about bigger-picture strategy and organisational matters.
As a result, for the first time in many years, my headspace has been mostly focused on the work in front of me. I.e. – how to understand and minimise social and human rights risks from technology, in particular online platforms, search engines, and AI. And, of course, 2024 raised a lot of difficult political situations, which make 2025 and onwards feel very uncertain. So my reflections this year mix these three aspects: fully re-integrating back into organisational work, tech policy, and the political situation of 2024 -2025.
I’ll arrange the reflections around a few key words: Elections, Impunity, Networks, and Random Wisdom. They start quite depressing, and quite broad, but get somewhat more positive (or at least more practical) as they go on.
Elections
2024 was heralded as the “year of elections”. Half the world lives in countries which had elections in 2024 (but not, as was often annoyingly reported, that “half the world will be going to the polls” – which assumes 100% turnout of the entire population including, erm, children). The USA was obviously the main event, but also India, the European Parliament, the UK, France (unexpectedly), and many others. I wouldn’t say (m)any had massively surprising outcomes, even the Trump comeback – on a quick check, in late 2023/24 he was still doing well in polling, betting markets, and forecasting. Maybe Narendra Modi doing expectedly badly in India was a surprise (and I don’t know what implications it has), but he still won. The left unexpectedly outperformed the right in France, but the expected chaos still reigns.
Overall, the main trend was “incumbent parties getting a kicking”. Another trend seems to be a drift towards voting for the cultural right, in particular regarding immigration and climate; but it’s hard to work out how much that’s a side effect of the “kicking incumbents” trend largely hurting centrist/centre-left parties, versus showing a genuine organic change in public opinion. It’s probably both, but working out the ratios will be an important task for progressive politics going forward.
For my work, the obvious implication of this “year of elections” was how social media would feature in all that. Usually this turns quickly to the topic of disinformation (or misinformation, or malinformation, or whatever). My long-held view is that dis/mis/malinformation are problems (this paper does a good summary), as part of wider poisonous political rhetoric, not as the main problem. But electoral events have a habit of focusing discussions on bad ideas like “votes get swayed by online fake news” and “social media is the main cause of bad stuff”. I think this lets lots of other actors off the hook, and also risks dismissing genuine democratic issues. I also think disinformation, while a problem, gets way disproportionate attention to other issues.[1] Given the rapid adoption of Generative AI in 2023-24, concerns around “Deep Fakes” looked like they could exacerbate that. So I went into 2024 bracing myself for a lot of annoying discussions.
But as 2024 unfolded, I was pleasantly surprised. I did not feel the year in tech policy was completely dominated by election monitoring. It featured in, but did not overtake, other important policy discussions – like data access for research, or assessing the impacts of AI. Where elections were discussed, there were good – if a few years late – pushbacks against the emphasis on fake news (and against the AI-driven “Deep Fake Apocalypse”), and also against jumping to “foreign interference from Russia” arguments rather than home-grown domestic bad stuff (AKA “the call is coming from inside the house”).
Of the elections I know about, I still cannot point to a single one in which I believe social media was clearly decisive – except perhaps the cancelled Romanian one at the end of this year. Social media has always interacted with the (still very powerful) mainstream media. When we hear “people get their news from social media”, it has still often been news from mainstream outlets (at least, as part of a mix). Trump is, to me, a TV politician; in 2016 his tweets, like his rallies, were broadcast via TV reportage from friend and foe alike (see of my favourite satirical lines).
But we still need to pay attention to the more diffuse impacts of social media, outside of elections: a general discourse in which hatred for the other side is normalised, and pressure for politicians feel to constantly make instant, public-image-focused decisions about everything, rather than spend time being strategic. I also think that future elections – even the next cycle – could be more dominated by social media, as we increasingly see people (especially young people) getting and distributing news and opinion purely via social media. And while I think “deep fakes” are overblown, how can trends towards AI summarisation – which aims for “one answer”, and makes untransparent choices about what to focus on – address something like political debate? I think such questions will be a big focus of my work next year.
And finally, the outcomes of the elections this year raise various concerns for people in my line of work. A major one is that research into social media is directly threatened by the new US administration, largely through intimidation by investigations and lawfare; this has been covered well elsewhere. A shout-out to the Coalition for Independent Tech Researchers who are doing great support work at these times. In the West this is largely a US-centred problem right now, though we have seen some similar warning signs in Europe. But for what I’m working on right now (which, very broadly, is questions of how to detect / create accountability for / mitigate tech-enabled risks and harms), the word that’s going through my mind as I think about tech & politics going into 2025 is…
Impunity
We already saw this in 2024. Most obviously from Elon Musk, who is somehow still seen as a free speech champion despite incredibly anti-free-speech actions on behalf of the Indian and Turkish governments. But also from Zuckerberg and Meta, who closed down Crowdtangle, a tool used by researchers and journalists to monitor trends on Meta platforms, shortly before the US elections (with no good reason); and have also basically absented themselves from engaging with politics. Also Sam Altman removing the board which provided oversight of OpenAI, and so on, and so on… largely with no bad consequences for any of them. (The one exception may be Telegram now actually doing basic moderation stuff after their CEO was arrested in France for not cooperating with law enforcement).
I expect the outcome of the US election will hypercharge this sense of impunity. In particular, Musk is obviously not going to follow the EU’s Digital Services Act rules for social media platforms, probably quite theatrically and petulantly (I’m imagining him posting a 10-Euro note in response to any fines). The EU Commission will eventually have to make a decision about whether they can and will ban X in the EU. This decision may not be forced in 2025, but kicking the can down the road will just make the whole Digital Services Act look toothless.
But there are deeper questions to grapple with beyond the impunity of tech CEOs (important as that is). There’s now a very broader coalition who are very sceptical of regulation, running from the unlikely bedfellows of the MAGA movement in the USA to Mario Draghi and supporters in the EU. There is an increasingly pervasive narrative that “regulation” (in general) in the EU is part of its economic stagnation. This debate is heated, but worth taking seriously amongst supporters of regulation: given that narratives like these can also drive worrying political developments - see e.g. the impact of a poorly-communicated and rolled out heating regulation in Germany on the growing support for the far-right AfD party.
The mere idea of who is “pro” and “anti” regulation is itself part of the fight. It’s a running joke that everyone (including CEOs) says “I’m in favour of regulation, but good regulations”, before defining “good” as “whatever supports me”. But it’s a question I want to treat as more than a joke this year. I want to think about it more seriously, in various ways: re-engaging with questions of how to regulate, including ideas of “creative bureaucracy”; looking into historical case studies; and the role of communicating regulation well (I think the EU has a habit of making regulations sound stronger than they actually are, which then frightens everyone; and then making the actual details unnecessarily hard to find).
But also – I’m not a regulator. My interest is in impunity and accountability, not regulation per se. If there’s other ways to get there, besides regulation – I’m interested. But I think neither shame nor appeals to morals can be relied on. I’m not anti-capitalism, but I think reliance on market solutions brings harms, especially to vulnerable people (see the impacts of climate change on poorer countries, or the traumatisation of people in Kenya to cheaply detoxify AI models). But this year, and ahead, will require those of us interested in accountability to think creatively.
Ok, those were some negative things. What am I more positive about?
Networks
Last year I wrote about:
the wider [EU] tech-policy community … who are proving an eminently driven, helpful, and enjoyable group to work with. I look forward to more in 2024.
I maintain this view. I believe that, when faced with the resources (and sometimes impunity, as above) of large companies and hostile political actors, leveraging distributed networks is key to getting stuff done. It also appeals to me as someone who enjoys, and has a strong tendency towards, building connections with people (to an extent that has become a running joke amongst my friends).
With such a range of tech policy topics, it’s been extremely helpful to have a variety of different expertise and experiences a quick Signal message away – even if that’s just to say “hey, are you also utterly baffled by [latest thing]”? My Head of Tech Research role also gave me more impetus to (re)-connect with plenty of researchers too, which is often fascinating. Also the interactions are often FUN, which helps a lot at difficult times. (Even if in-person events are usually in Brussels – which is not my favourite city, and German trains there are a level of unreliable between “3 hours delay” to “I end up having a spontaneous weekend in the Netherlands”. Entirely true story).
However – and here we return to a theme of my 2022 piece – the question of how to divide attention is always a challenge. This is partly at an individual level; how to keep up enough with the right people at the right time. But also what we might call the collective attention. There have been a few “flashpoints” around which the field has coalesced in this year, particularly around milestones in the Digital Services Act, the AI Act, and European Parliament Elections. This leads to helpful moments of collaboration and coordination. But it’s a difficult task to ignore these events – they are major, and it’d feel weird to be any EU tech policy organisation with nothing to say about these things – but it does make me wonder if too many eyes are sometimes pointing in the same direction. Where are the gaps, or the novelties that we might be missing?
So part of my 2025 will be thinking about the unusual and novel connections I could build between me (as an EU tech policy NGO person) and other spaces. I think a lot of my old employer CASM Technology, and their use of Natural Language Processing to analyse social media harms. It was a very cool and productive approach, which emerged from quite serendipitous meetings between political think-tankers and academic Natural Language Processing experts. So next year I might spend more time hanging around hacker-esque spaces like Kaggle or the Chaos Computer Club. But I don’t just want that to be tech-focused. As I mentioned above, I think there’s a super interesting role for historians of technology / regulation / the media to play in current debates. I’ve also mentioned in a few quarters that I’d love to see more anthropological work around how tech companies, developers, and risk assessors operate (how to get access to such research subjects is another question...).
Maybe there’s people making those connections already, and I’ve just missed them. If so, get in touch!
Random wisdom
Finally - going fully back into an organisation reminded me of random organisational wisdom I’ve been handed over the years. A few are below. (A lot of these fall into the “that’s obvious” category, but different ones will be obvious to different people).
Strategy is choosing what not to do.
Stuff rarely gets done without deadlines. Artificially imposed deadlines, e.g. through a process like OKRs, can help get the “we should do this thing sometime” tasks actually done. (But also – if everything is an urgent and
important deadline, worth questioning strategy).
Learning to iterate is super helpful – i.e. build things piecemeal, with lots of feedback, rather than labour away until a big reveal. Hoarding work for too long until it’s “ready” is rarely helpful, nor is only having a few major “feedback moments” and no testing in between those. There is, of course, an important element of skill to be learned; asking for someone’s attention has to be done in a way which minimises the mental load on them. But generally I think that’s about how one asks for feedback, not how often.
(There is also a link here to tech policy: one of my biggest frustrations with EU consultations is that we get various opportunities to discuss what we want in the abstract, but then generally long waits followed by just one chance to comment on actual legal texts. And then it’s locked in).
Somewhat related to iteration, a thing I learned from the civil service. People, particularly senior people, don’t want “more information” or “more background” to make their decisionmaking even more complicated. They want proposals and suggestions that they can answer “yes” or “no” or “I need more info about x”. They may reject them, it may turn out you misunderstood their brief. But I still think “here’s what I think you want” opens up better discussions than “let’s discuss the request”.
Delaying your less important meetings, time to discuss broader issues, check-ins etc. in favour of urgent / high-priority stuff is merely a quick fix which just saves up problems for the future.
So - Happy New Year
I hope this was an interesting, maybe useful, and not too depressing look back and ahead. At such times, one thing I take positivity from is just seeing good work being done. Many of you reading this do great work in politics, in solving social problems, in human rights… And even if I’m not involved, or they don’t affect me, it still gives me energy and hope to see it happen. So keep going, and happy new year.
[1] On the disproportionate disinformation gets: Compare the EU’s Code of Practice on Disinformation - now on its second version with 44 pages and over 40 signatories – with the Code of Practice on Hate Speech, which is still using the 3-page first version from 2018 and has 12 signatories.
Fun Fact About: Angela Merkel
I haven’t read Merkel’s memoirs yet; I’m currently slowly making my way through a biography of Willy Brandt (in German, as English ones seem weirdly lacking). But the reviews have not been good, and it seems Jeremy Cliffe’s very good critical piece about her, just before she stepped down, remains quite prescient. He lights on her phrase “driving by sight” as extolling her tactical skills of circumnavigating obstacles, without a clear strategy of where she was driving to.
I have no disagreement with these criticisms of Merkel. I will say, from a couple of people I know who have met her, she is apparently she’s much funnier in person than you’d expect and does quite good impressions of other politicians. From the reviews, I also found this story particularly entertaining (from Politico):
A famed photo of Merkel and former U.S. President Barack Obama, with whom she had good relations, on the sidelines of the G7 summit in Germany in 2015 shows Merkel with her arms outstretched, gesturing animatedly. Many have speculated about the subject of their discussion. But according to Merkel, she was merely recounting the size of a beach chair.
Recommendations
First, a request for recommendations: As lots of terrible things are happening in the world, I’d be happy to share info for charities that you can recommend which reliably do good, impactful work. Across the world the situation in Gaza continues to be horrific, but other situations like those in Sudan receive less attention. And the political situation in many richer countries will also lead to some pressing needs for help.
Books: In 2024 I really liked Free: Coming of Age at the End of History, the childhood memories of Lea Ypi, an Albanian political scientist at the London School of Economics, who was 11 when Communism ended in Albania. I’m not normally into sci-fi, but This Is How You Lose The Time War was very creative and good fun. The short stories in The Visiting Privilege by Joy Williams are extremely melancholic, a really great writer building subtly but powerfully emotional pictures. I’m also trying to improve my cooking and really enjoying the tips and things to try in Salt, Fat, Acid, Heat.
Trips: Of random places I visited this year, I’d recommend Slovenia (particularly Ljubljana), Strasbourg, Freiburg in southern Germany, and Hamburg & Lübeck. If you happen to be going to Colombia, I’d strongly recommend Cali and Salento over Bogotá (though the latter has some good museums). As I’m increasingly discovering that most of my friends who’ve visited England have only gone to London, I’d also recommend Chester and York for historic sites, and Manchester for fun.
Games: Maybe a bit late for gift-giving, but for next year if you have board or card games fans to buy for, I’d suggest perusing the Spiels des Jahres (Game of the Year) winners list running back to 1980. Although a German award, it is renowned worldwide as one of – if not the – top games awards.
Non-recommendations: Over Christmas I read the Financial Times’ Business Book of the Year Supremacy: AI, ChatGPT and the Race That Will Change the World; and watched the much-hyped film The Lives of Others about the East German Stasi. Both underwhelmed; but happy to discuss if you think I’m missing something!
Funnies: As this piece was quite depressing about social media, here are a few posts I’ve found amusing from across a range of platforms – from Twitter (but archived on reddit) to Instagram to Bluesky and beyond. THIS is the internet worth protecting.
Thanks for reading. Please do share with others using this link, and let me know your thoughts via this short poll.
Comentários