Malte Baader

In two studies using AI-assisted qualitative interviews, we ask a representative sample of the German electorate (N = 1, 013) to reflect on the most important societal events of the last 15 years. We document that far-right supporters construct systematically different narratives about the recent past. First, they recall different events-with a strong focus on migration-related issues. Second, even when they do consider the same events, they are more likely to blame the establishment and describe events more negatively and with greater emotional intensity. We further document that these narratives are associated with policy preferences, and that farright supporters are more inclined than supporters of other parties to prioritize cultural over economic issues. Together, our findings reveal that partisan divides extend to how citizens remember and interpret shared national experiences.

Digital transformation creates opportunities for organizations but also raises ethical risks related to privacy, surveillance, discrimination, and data use. Corporate digital responsibility (CDR) emphasizes the need to support responsible decision-making when evaluating digital initiatives. One widely used approach involves checking questions that encourage ethical reflection, yet little is known about their effectiveness. Across two experimental studies with professionals involved in data-related activities, we examined two types of checking questions: Ought-Questions, which prompt comparisons with normative standards, and Is-Questions, which direct attention to project-specific ethical risks. Contrary to expectations, Ought-Questions did not reliably reduce the acceptability of ethically problematic digital projects and, under some conditions, produced a boomerang effect. These effects depended on participants’ recognition of normative discrepancies and ethical risks. In contrast, Is-Questions, alone or combined with Ought-Questions, consistently promoted more critical evaluations. The findings show that ethical decision aids can both support and undermine responsible judgments in digital contexts.

AI decision assistants are entering domains where moral judgments matter, yet their effects on ethical choices remain poorly understood. In a preregistered multi-wave experiment, participants faced repeated ethical dilemmas with or without an AI assistant that scaffolded reflection without recommending actions. AI assistance increased decision confidence (Cohen's d = 0.40) more strongly than it increased ethical choices (d = 0.21), creating a confidence trap. The pattern persisted when controlling for choice improvements (residualized d = 0.36), and deeper reflection predicted choices but not confidence, consistent with confidence responding to the procedure rather than its content. Effects were strongest among morally motivated participants. These findings indicate an AI governance risk: decision support may inflate users' confidence more reliably than it improves stated-choice quality, and evaluations should not treat user confidence as evidence of effectiveness.

Unethical behavior, deception, and fraud are major concerns in corporate governance. This paper examines the effectiveness of contemplation questions (CQs) as ethical decision aids intended to activate moral agency and prompt employees to consider their actions from various perspectives (e.g., self, peers, company). Although we find that 44% of S&P 200 and Fortune Global 200 companies use CQs, their behavioral impact is not well understood. Using two preregistered, incentivized vignette experiments, we show that the effects of CQs depend critically on the level of engagement they require. In Study 1 (N = 1,986), where participants were only passively presented with CQs, we find no impact on ethical decisions, either in the aggregate or by reference group. In Study 2 (N = 1,322), where participants were actively engaged by answering CQs, we again find no overall effect, but uncover a striking heterogeneity: CQs marginally increase ethical decisions among individuals with high moral identity, yet significantly reduce such decisions among those with low moral identity. These findings align with a conceptual framework of motivated moral reasoning and suggest that while CQs can benefit some individuals, they may backfire precisely among those predisposed to unethical behavior.