How can we maintain solidarity and strengthen our democracy in the age of AI? This question ran through all three days of re:publica 2026 – and was also reflected in the four panels in which the bidt participated. These addressed a wide range of challenges posed by digital transformation: from communications surveillance and social media bans, through class concepts in the context of data work, to the role of digital technologies in the documentation of wars.
What is society’s stance on state surveillance of communications?
Under the title “Never gonna give up your phone?”, Professor Sabine Pfeiffer, Chair of the bidt Executive Board, and Bruno Albert, Research Officer, presented findings on public perceptions of the controversially debated state surveillance of communications in Germany.
Pfeiffer and Albert analysed the debate across six key dimensions – including the exploitation of security vulnerabilities, the use of AI, institutional control, and concerns about mass surveillance. Their analysis was based on a vignette study involving 1,250 respondents as part of the interdisciplinary research project “State Interventions in Private End Devices for Law Enforcement” (SEpES).
© bidt
The results show that support for surveillance measures depends heavily on clear constitutional limits. Judicial orders, restricted access to data and a limited scope of the measures are particularly important. Blanket or far-reaching access, on the other hand, is met with clear rejection.
The use of AI in communications surveillance was also assessed in a nuanced manner: whilst AI is viewed with scepticism when used to analyse surveillance data, the assessment is more positive when AI serves to protect privacy.
The conclusion of the presentation: Proportionality and the protection of fundamental rights are decisive for social acceptance – rather than the specific technical implementation of the surveillance measures.
Further links
Ban or regulate? Why the social media ban is polarising
The social media ban for under-16s, which has been in force in Australia since December 2025, is also sparking debate in Germany – particularly regarding the extent to which such a ban would make sense here and whether it could truly protect young people from the challenges of social media.
These questions were addressed at re:publica 2026 by the panel “I forbid you to do that!!1!1! – Australia’s social media ban for under-16s: hard facts rather than gut feelings”. Moderated by Salwa Houmsi, head of think tank Dr Roland A. Stürz from the bidt discussed the initial findings from Australia and what Germany can learn from them with Dr Josephine B. Schmitt from the Centre for Advanced Internet Studies (CAIS) and Dr Hannes-Vincent Krause from the Weizenbaum Institute.
© bidt
The focus was particularly on the question of scientific evidence. Stürz emphasised that a distinction must be made between scientific analysis and socio-political assessment. Research can highlight the risks and opportunities associated with social media use – yet clear causal links have so far been difficult to prove. Stürz also emphasised that political decisions cannot be answered by science alone. Social values and the need for protection also play a role; flexible regulation with the possibility of subsequent adjustments is therefore crucial.
Psychologist Hannes-Vincent Krause described a potential blanket ban as “fear-driven knee-jerk action”. Research has so far shown neither conclusively that social media fundamentally causes mental health problems, nor that bans are actually effective. What matters is not so much the sheer amount of time spent on social media as the type of content to which young people are exposed.
Josephine B. Schmitt from the Centre for Advanced Internet Studies (CAIS) also warned against simplistic solutions. Social media are not merely spaces of risk, but equally places for identity formation, information and social participation. Bans could even make problematic dynamics more invisible – for example, through fake accounts or private groups. Instead, there is a need for stronger platform regulation, the consistent enforcement of existing rules such as the Digital Services Act, and greater investment in media literacy and youth work.
A central element of the panel was the so-called “Hot Seat”, where young people themselves had their say. They described social media as both a burden and an important social space. Furthermore, the teenagers and young adults criticised the fact that they are often overlooked in political debates.
© bidt
Despite differing positions, there was broad agreement on one point: the debate on social media in the context of youth protection requires greater nuance, more scientific evidence and greater media literacy rather than simplistic black-and-white solutions. Platforms must be held more accountable, without depriving young people of digital participation across the board.
Further links
Policy Brief (In German)
LinkedIn (In German)
Digital traces of war: How technologies make human rights violations visible
How can war crimes be documented when traditional reporting is absent, evidence is fleeting, and digital platforms delete or distort content? Doctoral students Lama Ranjous, a member of the bidt Graduate Centre, and Anastasiia Liulina explored this question in their presentation “Digital Technology and Conflict Documentation in Syria and Ukraine”. They are investigating how digital technologies, decentralised archives and forensic methods can help transform volatile data into reliable evidence of wartime events
It becomes clear right from the start: this is not just about technical procedures. Ranjous and Liulina met in person for the first time at re:publica 2026. Prior to that, they had collaborated exclusively online. However, both share experiences of war, displacement and loss. Ranjous conducts research at the bidt on practices for documenting human rights violations in conflict contexts and examines the opportunities and limitations that digital infrastructures present for this work. As a member of the bidt working group “Critical Digital Geographies”, she also explores how digital tools, data and platforms relate to human rights, critical cartography, justice and accountability.
© bidt
The lecture focused on the “politics of visibility”: whose suffering is seen, and whose suffering remains invisible? Ranjous and Liulina demonstrated that without documentation, war crimes not only remain invisible in the media but also become difficult to grasp legally and historically. Anything that lacks military, political or media relevance easily disappears from public perception. Digital human rights documentation seeks to address precisely this issue: “It’s the work of making the invisible visible” – for instance through witness statements, photos, videos, satellite imagery, flight data analysis and analyses of chains of command.
A look at Ukraine showed just how significant such digital traces can be. The Ukrainian woman Iryna Filkina was identified after her death, partly on the basis of her strikingly painted fingernails. The image of her hand became a global symbol of civilian suffering in war. Later, Ukrainian investigators named a Russian commander as a suspect in connection with her death. The case exemplifies how visual traces, social media, journalistic research and legal proceedings can intertwine.
The work of the Decentralised Damage Mapping Group, of which Ranjous is a member, also highlights this intersection of technology, documentation and accountability. The international initiative uses Earth observation data to systematically record and visualise damage in conflict and disaster zones.
The two researchers’ presentation made it clear: digital technologies do not document war in a neutral and objective manner. They play a decisive role in determining which acts of violence become visible, verifiable and negotiable.
Or, as Ranjous and Liulina put it in conclusion:
When you see the visible, ask what remains invisible.
Further links
Working Group
Graduate Center
Click meets Class: Who actually works for AI?
Artificial intelligence is regarded as a driver of innovation for the future – yet behind many AI systems lies invisible human labour. Professor Sabine Pfeiffer (bidt) and Dr Milagros Miceli (Weizenbaum Institute) analysed this tension in their lecture “Click meets Class: Technology, Power, and Labour Between Inequality and Solidarity”. The focus of the lecture: data work, digital exploitation and the question of who bears the costs of the AI boom.
© bidt
As early as 2021, the company OpenAI employed people in Kenya via external service providers to evaluate toxic training data for ChatGPT. Precarious employment contracts and high levels of mental strain coupled with low pay are the order of the day. Responsibility for working conditions is frequently outsourced.
The global research initiative “Data Workers’ Inquiry” brings precisely this often invisible work to light. After all, data work has long been happening everywhere: in content moderation, data annotation or algorithmic surveillance in the workplace. Whether copywriters, Uber employees or staff at major tech corporations – digital work is increasingly being measured, evaluated and optimised across all sectors. This has simultaneously brought a fundamental misunderstanding surrounding AI into focus:
These systems are by no means autonomous or error-free. AI does not truly understand the world; instead, it often relies on statistical patterns and, in some cases, outdated data, producing forecasts that are prone to error. This is precisely why we still need people with experience, contextual knowledge and sound judgement.
Prof. Dr. Sabine Pfeiffer To the profile
Studies also show that the promised productivity gains from using AI often fail to materialise. In one study, for instance, programmers working with AI support actually worked more slowly – yet believed themselves to be more productive. Other studies show that companies have so far seen hardly any measurable productivity gains from AI investments.
At the same time, AI is already having a massive impact on our work today. Time pressure, constant monitoring and cognitive overload are on the rise – a phenomenon now discussed under terms such as ‘AI Brain Fry’. This becomes particularly problematic where AI is used in sensitive sectors, such as medicine or transport. When technologies are developed without taking real-world working conditions into account, new risks and dependencies arise.
The question is therefore not only to what extent AI replaces human labour, but also under what conditions human labour makes AI possible in the first place – and who benefits from it. It remains to be clarified what responsibility society and companies in particular (should) bear. After all, behind every ‘smart’ technology, behind every AI chatbot, there are ultimately people.
Selected highlights
Research Focus
Research Project SEpES
Best-of on LinkedIn (In German)
Video recordings
By loading the video, you agree to YouTube’s privacy policy.
Learn more
Image gallery
© re:publica 2026, Berlin; Photo credit: bidt
