23 NOVEMBER 2026, 2:30pm - 4pm

Institute of Philosophy, Jilská 1, Prague 1, meeting room

Abstract:

Agentic AI, corporate surveillance, and incumbent state power could enable authoritarian power grabs and potentially irreversible value lock-in. While this does not require human-level intelligence, AI already matches or surpasses human performance in some crucial domains, worsening the threat. Public scrutiny is essential to democracy, yet frontier labs' internal capabilities greatly exceed what outsiders can access or observe. When governments ignore risks or pose misuse risks themselves, and executives cooperate or cave, frontier researchers' civic conscience and bargaining power may be among the last effective checks. However, their leverage is perishable: Frontier labs explicitly pursue automated AI research, which may soon erode researchers' bargaining power. This talk argues that frontier-lab workers should unionize and attempt to enforce industry-wide safety and security standards (via collective whistleblowing, targeted research strikes, etc.). The public should support their organizing—and press them to exercise this responsibility. Unions cannot replace democracy, but may help preserve its possibility in a time of peril. 

Adriano Mannino is Professor of Philosophy at the Munich University of Applied Sciences. He was previously a Technology & Human Rights Fellow at Harvard University, a Postdoctoral Fellow at UC Berkeley, and a Doctoral Fellow at LMU Munich. His interests are in normative ethics (aggregation, welfare, rights), applied ethics (AI, biomedicine, nonhuman welfare, catastrophic risks), and political theory (AI, democracy, public discourse).

Free entry.

This event is organized as part of the project Human-centred AI for a Sustainable and Adaptive Society (HumanAId, reg. no. CZ.02.01.01/00/23_025/0008691), implemented by the Institute of Philosophy CAS. The main objective of the project is to develop methodologies and tools that will enable the potential of large language models (LLMs) to be used in a way that is in line with the value and normative requirements of specific users in civil society and government.