Political Tech Summit 26, Berlin

These are my notes from Political Tech Summit 26. I have focused on those sessions I attended. The technology support presents what looks like AI written summaries of the sessions which were mainly panel sessions. I have quoted from these summaries and in one case sought to re-summarise using another LLM. I have tried to humanise the text so that the article remains in my voice but also informs you of what happened.

The first session I attended was entitled, “Protecting our Future: Cybersecurity for Society, State and Public Administration”. The panellists were Marie-Agnes Strack-Zimmermann MEP, Chair of the European Parliament’s defence committee and Wieland Holfelder, Vice President and Regional CTO for Google. The conference preview promised the argument that society has transitioned from a state of peace into a perpetual “cyber war” where digital attacks precede physical conflict and would conclude by agreeing that Europe must shift from mere regulation to “operational resilience” by leveraging AI to counter scale-based threats from global adversaries like China and Russia.

MSZ was clear that Europe need to own its solutions while not explicitly stating that US was a source of threat. She cited the recent Davos conference as a source of European concern and unity. Hofelder was explicitly asked, how, in a world of the US Cloud Act, Google could be expected to keep European citizens’ & governments’ private data private. He argued that their solution of third-party key holders under contract, who would need European incorporation and ownership, should be sufficient and they have made a lot of effort to present their solution set as fit for purpose. One problem is that while measures to encrypt data on the wire and at rest exist, data is unencrypted while being processed.

The panel did not spend much time examining the EU’s legislative defences, but the US consider the Digital Services Act to be against the interests of the big-tech companies. This is because if enforced it’s an essential part of European citizens’ defences against both hate speech and AI generated nonsense and lies.

The conference summary includes the paragraph,

Holfelder presented AI not just as a risk, but as the primary engine of defense. He noted that AI has revolutionized threat detection, stating, “In the past, you would need to find the needle in the haystack. Now, we just burn the haystack and find the needles.” Highlighting Google’s “CyberShield” technology, which increased defensive capabilities by over 300% and kept Ukrainian government websites operational, Holfelder argued that cloud-based solutions offer superior security to localized systems. However, this sparked a debate on sovereignty; while Holfelder pointed to Ukraine’s reliance on the cloud, Strack-Zimmermann cautioned that Europe must find its own path to avoid over-reliance on a U.S. administration that may no longer be a guaranteed ally, stressing the need for technical verification without blind trust.

I like Holfelder’s metaphor and for many/most of us relying on the talent of Google’s cybersecurity staff is a better answer than rolling our own. There remain questions about US companies needs to meet US compliance laws before those of their EU or host countries; this remains a risk for non-US companies and people.

Another thing that I was curious about, is that while the session was advertised as beyond the firewall, there was no mention of our historic ideas on cyber defence, cyber-security nor the international standards defining good practice, such as COBIT & ISO27001. We know a lot about keeping data secret & truthful, we shouldn’t forget it.

On the 2^nd Day, I attended a session entitled, “Beyond the Firewall: Online Safety, Regulation, and Why Defense Alone Is No Longer Enough”.  This was chaired by Damien Collins, who has an admirable record in UK Tech regulation; he is an ex-UK MP. He was joined by Alexandra Geese MEP,  Zoe McKenzie MP [Australia] and Lisa Steigertahl from Deloitte.

Participants have access to summaries of the session, and the summary for this includes the paragraphs,

… [They] highlighted a fundamental tension between “engagement-based ranking” and liberal democracy. Geese argued that algorithms designed to exploit rage and fear are “systemic risks to democracy” because they prioritize our “animal brains” over rational discourse. While Lisa Steigertahl from Deloitte … advocated for “AI sandboxes” to test safety by design, the political speakers emphasized that the issue is primarily one of political will.

… the [panel] contrasted Australia’s gradual regulatory path with the more confrontational tone surrounding EU-U.S. digital relations. McKenzie noted that Australia established its e-safety framework early, giving platforms time to adapt, whereas the EU now finds itself negotiating digital regulation amid broader disputes over security and trade.

Mckenzie also noted that the tension between Australia and the US Tech companies is almost certainly less than that between them and the EU due to the smaller population (and thus profitability). She also noted that the pre-teen ban on social media is popular amongst voters.

Geese’s statements about encouraging rage, reminded me of last year’s presentation by Beth Goldberg who also highlighted the strong linkage between profit and rage and the resulting amorality of the platform providers. Geese was particularly forthright on Grok/X’s use of AI to attack and humiliate woman and girls.

Their conclusion was that the EU has the enforcement laws it needs, the problem is the seeming lack of a will to enforce within the Commission, which is  fearful of US retaliation in the trade and defence spheres. They also argued that EU needs to have its own social media infrastructure to replace the profit driven US social media platforms. This and other contributions over the weekend reminded me of the growing calls for European IT sovereignty. See https://davelevy.info/wiki/it-sovereignty/

Damien Collins was an excellent moderator; it was great to see  he hasn’t given up.

Ned Mendez in  a presentation called “Signal Integrity and Audience Insight in the Post-Platform Era Findings” argues that twitter/x in particular, but also facebook have no-one that political activists want to talk to. These platforms are also ‘poisoned’ by amoral players who post for money while the platforms chase for profit at the expense of a decent user experience enables and accelerates this aspect of enshitification.  

Here’s a concise summary by Mistral of Ned Mendez’s argument about the “post-platform” era and its implications for political strategy. He argues,

… that we are entering a “post-platform” era where digital signals—once reliable for political strategy—are now “compromised.” Platforms like X are not neutral; they actively   and shape what is visible, measured, and analyzed[1]. Algorithms create fractured realities: right-leaning users see reinforcing content, while left-leaning users face hostile, competing messages. High engagement metrics often reflect what platforms show, not what people truly believe.

In this world, local grievances can be “supersized” by international actors often working for money, creating “zombie signals”—amplified distortions that appear as national crises but are actually commercial noise. Reacting to these by politicians and the state undermines democratic resilience.

I paraphrase Mistral & Mendez by saying that the answer may be to move away from “elite bubbles” (MPs, journalists, activists) to “non-politically aligned environments” (everyday spaces like hobby forums or local community pages). If a political issue appears in a gaming forum, or on a football terrace, it signals genuine public concern. I describe that to myself as having a control sample.

If the meme’s are not getting traction, we can conclude that not every social media wave deserves a response. Success in the future will depend on distinguishing real signals from noise and ignoring profit seeking big tech nonsense.  This aspect of the talk reminded me of Labour’s nineties instant rebuttal solution, and wonder if their time has passed, certainly no-one is talking about that this year.

The final session I in this report was called, “Is Social Media dead?”. The panel members were Ben Guerin, co-founder of Topham Guerin and Clare O’Donoghue Velikic, Founder of ODV Digital.

They argued that Social Media is changing, it’s losing its intimacy and the platforms are  becoming media presences i.e. it’s losing the “social”. They also argued that there is a move from text to video, and that these need to be short. Tik-tok is now more viewed than Google although I wonder if AI will help restore[2] a balance. On a personal note, and I was the only person admitting to posting to Facebook in the last 24 hours, I prefer text as I can control the time I spend and make my own priorities by skipping the boring bits, but the panel argue that political commentators need to do [short] video first.

Velikic also argues echoing Mendez, see above, that the number of useful correspondents on the platforms is limited. In fact, she argued that it’s so limited that even advertisers should think twice.

From the conference summary, ‘Velikic argues that,

the best response is not to compete in a “race to the bottom” metrics-wise, but to be the “green grocer”—a voice of authenticity, humility, and humanity that provides a “beacon” in a sea of AI slop. There was a consensus that the “TikTokification” of society is leading toward a “post-literate” world where a politician’s entire platform must be condensed into a single, memorable sentence to survive the 30-second attention span.’

They make the point that rage gets clicks but is that what politicians want? Surely they want their ideas to convince.

The algorithms have not gone away and they are still opimised for profit.

Again from the summary,

The bottom-line is that while social media as we knew it is dead, it has been reborn as a multimodal, video-first battlefield where “speed will decide the truth”. To win in 2026, political actors must stop throwing money at Zuckerberg and start harnessing their own “digital civilian patrols”—engaged supporters who act in real-time to defend narratives on messaging platforms. As Velikic concluded, the only way to beat the machines is to show the algorithms that people actually want “real humor, humility, and humanity” over automated slop.

My summary conclusions from the conference are that

• My view that, very few politicians or their technologists or parties know how to persuade, remains true. A lot of the political technology today focuses on GOTV and fund raising.
• European politicians are angry about Grok’s nudification tools and there is some resolve to ensure that US companies don’t make money by proselytising hate.
• The platforms are no longer useful as means of contacting voters, and the arguably no longer useful as an advertising platform.
• Chatterton’s lessons from last year, that velocity and authenticity count more than a high volume/low hit rate communication strategy.
• The conference, or at least the bits I saw, were more technical and less political than 2025.

[1] Oh dear. Mistral writes with US spelling.

[2] No reason why it should, AI’s can make video now.