Trusting the use of AI by the authorities, the Swiss terms and conditions at the forefront

It is generally accepted that trusting authorities plays a key role in citizens' behaviour. Moreover, constitutional law places a duty on authorities to act in good faith and to behave in a trustworthy manner. At the same time, authorities wish to implement new technologies and AI systems to provide better public services and hope for efficiency gains. Swiss authorities were among the first to implement chatbots as AI systems. To try to limit their liability concerning the use of these systems, authorities on all levels of government chose to publish terms and conditions to advise the public and curtail the use of these chatbots. Although the legal qualification of these texts and the validity of their content are questionable, our research analyses the content and wording of these terms and conditions and the broad disclaimers they contain to determine their probable impact on trust. We argue that by refusing any accountability, the authorities present these services, and themselves by extension, as untrustworthy. This, in turn, might have a broader impact on the perception of AI in public services and the overall development of e-government services.