By Josef Lentsch, Founder of Partisan, the Political Tech Summit in Berlin, the Political Communication Summit in Vienna, was one of the Founding Board Member of NEOS in Austria. Josef is a lifelong political entrepreneur who first ran for class president at the age of 10. Since then he has co-founded several start-ups, including a political party, advised many political organisations and written a book on how others can innovate politics. A trained psychologist and public administrator, he likes to keep an eye on both the individual and systemic levels of politics.

On Invitation for LI’s Reimagined Newsletter – The Liberal Dispatch

Sign Up for The Liberal Dispatch

AI can help liberals listen better, work faster and make complex ideas accessible.

Artificial intelligence has arrived in politics. The public debate has gone straight to deepfakes, disinformation and dystopia. Fair enough: the risks are real.

Meanwhile, a quieter revolution is taking place inside campaign headquarters. AI is making political communication cheaper, faster and smarter across all stages – listening, researching, writing, translating, testing, responding and managing.

This matters for liberal parties. They tend to be smaller than their competitors, short of money and staff, and fond of explaining things at length. AI can give a modest campaign something resembling a research department, translation service, creative studio and rapid-response unit, all at once.

Inside the Campaign Machine

AI can summarise hundreds of conversations with voters and identify recurring concerns. It can turn a manifesto into a sourced question-and-answer tool. It can prepare a candidate for an interview, translate campaign materials, help volunteers answer policy questions, and turn a town hall into clips, summaries and follow-up messages.

It can also check whether the party is saying roughly the same thing across speeches, social media, leaflets and constituencies. Anyone who has worked in a campaign knows how difficult that can be.

Technology, however, needs direction. What is the problem? Who are we trying to reach? What do we want them to understand, feel and do? Without clear answers, AI simply helps a campaign produce confusion at greater speed. To paraphrase an old saying: a fool with an AI-driven tool is still a fool.

The Dutch liberal party D66 showed what a more thoughtful approach can look like during the 2025 election. It openly used AI to simplify the language of its manifesto and make its ideas accessible to more citizens. Alongside the full programme, it published a summary, an accessible-language version, a sign-language version and an English translation.

The wider campaign brought these ideas together under Rob Jetten’s optimistic proposition, Het kan wél: “It can be done.” Policy substance, accessibility and a clear story reinforced each other.

Liberals need to tell Stories

There is a lesson here for liberals. They like to explain. They believe that if people understand the policy, they will support it – which, it turns out, is not always the case.

Politics also requires a story: a recognisable conflict, a sense of direction, emotional clarity and a reason to care. Populists understand this very well. They repeat – again and again. They adapt their message to the platform. They make clear who belongs, who is to blame and whose side they are on.

As the UK-based consultant Jonathan Tanner recently wrote: “The magic isn’t in the technology—it’s in using technology to perfect the basics.”

AI will strengthen those basics across the political spectrum. For liberals, it can simplify language, structure choices and adapt information for different audiences. It cannot decide what matters. Political leaders still have to provide meaning, judgment and presence.

A pragmatic approach should distinguish between campaign operations, voter engagement and deception. AI can already improve research, coordination, translation, community management and content production. Personalised voter communication requires safeguards. Fabricated events, AI-generated hallucinations and abusive use of voter data require clear boundaries. Someone human must remain responsible for every political message.

The Next Election Laboratories

The next 12 to 18 months will be very informative. The US midterms in November 2026 will become a huge laboratory for AI-assisted fundraising, creative testing, voter contact, opposition research and rapid response. We will see conversational agents, synthetic video and increasingly autonomous campaign workflows.

France’s 2027 presidential election will be Europe’s next major test. Its highly personalised political culture will accelerate AI-generated video, real-time narrative analysis, digital surrogates and direct candidate-to-voter communication. It will also test how much speed and personalisation voters will accept before trust begins to break.

AI is changing where people encounter politics, too. Increasingly, voters are asking chatbots and search assistants what a party stands for. Parties need to make their positions clear, sourced, accessible and machine-readable. These systems are becoming a new group of political intermediaries. Liberals should start treating them accordingly.

AI gives liberals a chance to make politics more accessible, responsive, multilingual and participatory. Whether they take it will depend on political clarity, organisational discipline and the courage to tell better stories.

These are exactly the issues we will debate at Europe’s first Political Communication Summit in Vienna on 6–7 October 2026. The next campaign cycle will favour political movements that combine machine speed with human judgment—and give people a story worth joining.