Cognitive aids: Monitoring their performance, gauging their affective consequences

• Applicant: Patrick Weis, Ph.D.
• Subject Area: General, Cognitive and Mathematical Psychology
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 463896411

We use computers instead of mental arithmetic, translate text via AI chatbots instead of probing our own language skills, navigate with an app—possibly even in our home city. To understand everyday cognition, it is indispensable to also understand cognitive aid use. And while cognitive aid use is already important today, it will be pivotal tomorrow. However, our understanding of cognitive aid use is still developing. Problem solvers frequently prefer one aid over another based on superior performance. Accordingly, performance monitoring allows making informed decisions whether to use an aid. But previous research suggests proficient monitoring only in some, and imprecise or biased monitoring in other cases. What is at the roots of such imprecisions and biases? In Working Package 1 (WP1), I will reveal the underpinnings of a specific monitoring bias identified in the previous funding period: the primacy bias. The primacy bias indicates that problem solvers prefer cognitive aids that had performed well on initial encounters, irrespective of later performance decrements. Analogously, it indicates that problem solvers dismiss aids that had been slow initially even if their performance has since substantially improved over time. In WP1, I will disentangle the contributions of a passive attention decrement, actively boosted attention during initial episodes, and attention-independent estimation biases like anchoring to the primacy bias. Moreover, I will investigate temporal stability of the bias. Another neglected aspect is the affective footprint of cognitive aid use, which I will explore in WP2. Remediating the neglect is important since affect is tied to wellbeing, mental health, implicit metacognition, as well as reinforcement learning. Specifically, I will explore three potential affective consequences of cognitive aid use. First, problem solvers who use cognitive aids might miss out on positive affect which is elicited by finding solutions on one’s own. Second, problem solvers might be avoiding negative affect caused by the cognitive demand and cognitive conflicts that occur during unsupported problem solving. Third, problem solvers might experience negative affect during aid use. Receiving help from humans can trigger self-devaluation, including fear of appearing incompetent or feelings of inferiority, which could generalize to settings with non-human helpers. In concert, WP1 and WP2 have the potential to create insights that could both guide problem solvers in their cognitive aid use and also help designers in creating cognitive aids that are tailored to human capabilities and needs. By doing so, the proposed research will hopefully inspire further translational research on a topic of high societal relevance.

• DFG Programme: Research Grants