Abstract Empirical evidence has demonstrated the power of AI to influence human decisions and the risk of humans acquiring AI biases. Therefore, there is a clear need to develop strategies to mitigate such threat. In three experiments, set in a medical context, we tested whether warning individuals about AI biases and errors could mitigate the negative impact of AI biases on their decisions and reduce the transmission of AI biases to humans. In Experiment 1, participants received explicit information about the percentage of erroneous AI recommendations but with two different framings: in terms of AI accuracy or AI risk of error. Our results showed that emphasising the risk of AI errors, more than its accuracy, reduced people’s tendency to follow incorrect AI suggestions and to acquire biases from AI. In Experiment 2, a more general warning message alerting of possible AI errors and biases was also effective in reducing bias acquisition. Experiment 3 showed that, although the warning message provided some protection against bias, participants who received AI support still made more errors than participants who completed the classification task without any assistance. Experiments 2 and 3 also investigated whether the type of error made by the AI, a false positive or a false negative, influenced participants’ tendency to adhere to its suggestions, and the effect of the warning message. However, no significant effects were found. Overall, our results highlight the importance of informing users about the risk of AI error rather than focusing solely on accuracy.
Transcranial photobiomodulation (tPBM) has been employed for cognitive enhancement in healthy individuals. This study aimed to investigate the effects of tPBM applied bilaterally over the dorsolateral prefrontal cortex (DLPFC) on convergent thinking (CT), divergent thinking (DT), and the Stroop test. Additionally, we explored whether Stroop performance mediates the effect of tPBM on creativity. In this double-blind, between-subjects study, 56 healthy participants were randomly assigned to either the tPBM or sham group. tPBM was administered using near-infrared light (810 nm, 40 Hz; 50% duty cycle) over the right and left DLPFC for 20 min. Creativity was assessed at baseline and during stimulation using the Unusual Uses (UU) and Picture Completion (PC) for DT, and the Remote Associates Test (RAT) for CT, and the Stroop test. ANCOVA, controlling for baseline scores, revealed that the tPBM group scored significantly higher than the sham group on the RAT (F = 6.15, p = 0.016) and Stroop (F = 4.89, p = 0.031). However, no significant differences were observed for DT. The findings suggest that tPBM may be effective in enhancing CT, but its effect does not appear to be mediated by improvements in Stroop performance. These results indicate that tPBM could be a promising tool for cognitive enhancement in the healthy population.
The alignment of self-assessment judgements to formal grading (i.e., accuracy) is an essential process by which students calibrate their thinking about work quality to disciplinary standards. However, extensive research shows that students are not well-calibrated in their self-assessments. Relatively little is known about how the realism of self-assessment affects positive or negative emotions and self-efficacy. In this study, we examined the relationships between self-assessment accuracy (over- or under-estimation) and students' emotions and self-efficacy. A total of 112 higher education students wrote an essay, self-assessed their performance, received feedback, self-assessed again, and subsequently received the tutor grades for their essays, creating a potential discrepancy between self-assessment and tutor assessment grades. Based on the sequence of events in the experiment, we used autoregressive path modeling to examine the repeated measures relationship of emotions and self-efficacy over time. Starting values for the two emotions and self-efficacy were strong predictors across time. The discrepancy between tutor and self-assessed grades was introduced after two rounds of self-reported emotions and self-efficacy and regressed onto the third round of those variables. Students who overestimated their performance experienced a decline in positive emotions, an increase in negative emotions, and a decrease in self-efficacy when confronted with their inaccuracies. In contrast, those who had underestimated their performance had a concomitant increase in positive emotions and self-efficacy, and a decrease in negative emotions. This study highlights the importance of fostering self-assessment accuracy for emotional well-being and self-efficacy. Future research should delve deeper into processes that support calibrated realism in student self-assessment.
This paper evaluates the current maturity of automatic code-generation workflows for deploying modern CNN-based object detectors on embedded GPU platforms. We compare a native pipeline against a code generation pipeline through a Model-Based Engineering (MBE) approach, using YOLOv8/YOLOv9 inference on NVIDIA Jetson Orin Nano and Jetson AGX Orin as representative edge-GPU workloads. We report detection-quality metrics (mAP, PR curves) and system-level metrics (latency distribution and initialization overhead) under a controlled single-class scenario based on a CARLA-generated sequence with frame-level annotations. Absolute accuracy and latency values are scenario-dependent and may vary under different camera optics, illumination, motion blur, sensor noise, occlusion patterns, and multi-class scene. Results quantify the performance gap between code generation and native pipelines and show that, for the evaluated workloads, the automated pipeline remains less competitive in both latency and accuracy. We discuss the implications of this gap for deployment workflows in safety-oriented domains, and we outline bottlenecks that should be addressed. The study is intended as a controlled traffic-light detection micro-benchmark and does not aim to validate full ADAS perception stacks.
Background. It has been shown that addictions are stigmatized and that this stigma affects a person's well-being and treatment-seeking. This study aims to analyze whether individuals exhibiting various addictive behaviors-specifically alcohol use, drug use, gambling, gaming, and compulsive buying-experience greater levels of perceived stigma than those without such issues. Additionally, the study aims to examine differences in stigma based on sex. Methods. The sample consisted of a total of 136 participants from the general population. The participants' age ranged from 18 to 63 years (M-age = 42.13; SD = 9.32), and 78.7% were female. Results. There are significant differences in discrimination and disclosure between individuals who report problematic use of illegal substances and gaming and those who do not. For women, significant differences were found in discrimination and positive aspects subscale based on problematic use of illegal substance use and gaming. Among male population, there were significant differences between men with compulsive buying in discrimination and positive aspects compared to non-problematic use. Conclusion. Identifying the populations that are most susceptible to stigma is crucial for developing effective strategies for its prevention and reducing potential detrimental consequences.
