Join Michael Wesch and Elisa (EJ) Sobo for an exciting conversation on how to leverage generative AI for better teaching and learning! Together, they delved into the challenges presented by large language models and generative AI technologies, offering not only concrete suggestions for classroom practice but also timely reflections on the purpose of higher education.

"In meeting this new technology with creativity and purpose, we can reorient education’s compass needle back toward process as opposed to product – toward thinking about as opposed to merely recounting what others have said" – and as we do, we can bolster efforts to increase equity in higher education too: "The adaptations we create, not ChatGPT, could be the real revolution" (Sobo, 2023, 1). By guiding students to apply generative AI in innovative classroom activities and projects, we can empower our students to use this technology as a practical tool for inquiry and discovery while also interrogating its potential sociocultural impacts *(Sobo, E. (2023). Could ChatGPT Prompt a New Golden Age in Higher Education? Teaching and Learning Anthropology, 6(1). http://dx.doi.org/10.5070/T36160114)*



The REAL AI revolution: the benefits of Chatbot Augmented Instruction

Background
CSD programs are utilizing simulation as a pedagogical tool to support student learning and development of clinical skills. Simulation experiences are designed to allow students to practice clinical skills in a low-stakes environment and to increase standardization and equity across student learning experiences (Dudding et al., 2019; Quail et al., 2016). Another benefit of simulation is the ability to observe student performance to ensure professional competency and observe skills prior to students’ initial clinical placements (Macbean, Theodoros, Davidson & Hill, 2013). Simulation pedagogy includes various methods, but one prominent method is utilization of standardized patients (SPs) which introduces a high level of fidelity to simulation experiences given the SPs ability to respond and react to students (Zraick, 2020). SPs are trained to simulate a patient in a repeatable, standardized manner in order to ensure similar learning opportunities for all students (Dudding and Nottingham, 2018). 

Pacific University uses a community-based clinical model where students are placed with community supervisors for all practicum experiences. As such, clinical faculty work diligently to form and maintain relationships with speech-language pathologists (SLPs) in the community to develop, not only clinical placements but, meaningful learning experiences in the graduate program for students to gain “hands-on” experience before their first practicum in the spring semester. The Clinical Methods and Observation course (CSD 521), completed in the first-semester of the graduate speech-language pathology program, is a foundational training course for development of students’ clinical skills prior to external clinical placements in the local community. Historically, graduate students in CSD 521participate in active learning related to hearing screening procedures in a class lab. Students were introduced to the rationale for providing hearing screenings, procedure protocol, and how to provide results to patients or caregivers. In lab, students have practiced using audiometers to complete the hearing screening protocol one to two times with lab partners. Following this preparation, students were sent out to schools in the local community to assist the school’s SLP with hearing screenings. Generally this method of preparation was successful, but students commented that more structured practice and feedback prior to working with clients would be beneficial. 

Given this student feedback, the course instructors determined that inclusion of a high-fidelity, high-quality simulation allowing students to practice a hearing screening procedure with a standardized patient could bridge the gap from class to clinic. The instructors reviewed the current simulation evidence in communication disorders and sciences (CSD) and designed “perfect practice” opportunities for students to learn the hearing screening technical procedure. Following this practice, students were expected to participate in a simulation utilizing a SP to learn the nuance of working with a live client, tracking responses online, and managing client questions and behaviors. 

Purpose 
The purpose of this study was to investigate the performance and perception of graduate students following a simulated learning experience focused on administering and analyzing results of an adult hearing screening. Investigators worked to answer three main questions, including (1) How did students perform on the simulation, (2) What did students think about the simulation, and (3) Is there a relationship between how the student performed on the simulation and how they rated it?

Methods
This simulation was completed by 32 SLP students in their first semester of graduate school before students conducted hearing screenings in the community. The objectives were: (1) Introduce yourself/explain the task in clear terms without clinical jargon, (2) Complete a hearing screening following the appropriate procedure per protocol, (3) Demonstrate clinical/critical thinking skills throughout the encounter, and (4) Use an audiometer accurately and efficiently.

Prior to the simulation students were provided with pre-learning materials, the clinical educators reviewed the simulation objectives, and the hearing screening procedure was modeled. Students practiced with each other during a lab and open practice day with access to the audiometers and mannequins to practice with faculty present to answer questions. 

On simulation day, the supervising educator led students through a scripted pre-brief. The students individually administered a hearing screening to an adult SP. The supervising educators used a rubric to evaluate each student's clinical and communication skills. Once the simulation was completed, students debriefed with their supervising educators. The supervising educators led debrief session using the PEARLS method which provided students a structured outline for reflections related to the simulation, their feelings related to performance, and knowledge acquired from the simulation experience (Walter & Cheng, 2015). Following the debriefing session, students completed a Qualtrics survey to evaluate their performance and the simulation itself.


Results
Preliminary review of the data collected by clinical educator rubrics revealed the following:
- 81% (26/32) of students met criteria for explaining the task
- 81% (26/32) of students met criteria for following the hearing screening protocol
- 97% (31/32) of students met criteria for demonstrating critical/clinical thinking skills
- 66% (21/32) of students met criteria for using the audiometer efficiently

Common feedback for students from clinical educators was: ensure the client is facing away from the student clinician, apply headphones on correct ears, minimize jargon, and consider ergonomics for using audiometer while collecting data. 

Immediately following the escape room, students completed a Qualtrics surveyto provide feedback across several simulation design and implementation areas, including a) structure, b) receiving guidance and feedback, and c) debriefing about their learning. Data is yet to be analyzed, but preliminary review appears to show that students are confident that they are learning the content of the simulation and this simulation covers critical content to prepare them for clinical practice. Qualitative feedback from the survey indicated that students were happy to have the opportunity to make mistakes in a low-stakes environment, so they would be more successful with their first clinical experience.

Conclusions/Implications 
The outcomes of this study suggest that students benefit from practicing hearing screenings in a simulated context to increase clinical confidence prior to their initial community placements. This work has positive implications for student teaching and training, as well as potential for carryover to clinical contexts to improve hearing screening practical skills in early-career SLPs.


References

Dudding, C. D., Brown, D. K., Estis, J. M., Szymanski, C., Zraick, R., & Mormer, E. (2019). Best Practices in Healthcare Simulations in Communication Sciences and Disorders Guide. Counsel of Academic Programs in Communication Science and Disorders. Available from: https://members.capcsd.org/news/Details/capcsd-ebook-best-practices-in-healthcare-simulations-in-communication-sciences-and-disorders-guide-8147 


Dudding, C.C. & Nottingham, E. E. (2018). A national survey of simulation use in university programs in communication sciences and disorders. American Journal of Speech Language Pathology, 27(1), 71-81. https://doi.org/10.1044/2017_AJSLP-17-0015 


MacBean, N., Theodoros, D., Davidson, B., & Hill, A. E. (2013). Simulated learning environments in speech-language pathology: an Australian response. International journal of speech-language pathology, 15(3), 345–357. https://doi.org/10.3109/17549507.2013.779024 


Quail, M., Brundage, S. B., Spitalnick, J., Allen, P. J., & Beilby, J. (2016). Student self-reported communication skills, knowledge and confidence across standardized patient, virtual and traditional clinical learning environments. BMC Medical Education, 16(1), 73. https://doi.org/10.1186/s12909-016-0577-5 


Walter, E. & Cheng, A. (2015) Promoting Excellence and Reflective Learning in Simulation
(PEARLS), Simulation in Healthcare: The Journal of the Society for Simulation in Healthcare,
10(2), 106-115. https://www.doi.org10.1097/SIH.0000000000000072 


Zraick, Richard I. (2020) "Standardized Patients in Communication Sciences and Disorders: Past, Present and Future Directions," Teaching and Learning in Communication Sciences & Disorders, 4(3) https://doi.org/10.30707/TLCSD4.3/KHSI3441 



Using Simulation Practices to Refine Clinical Skills for Hearing Screenings

There is increasing pressure on social media companies to reduce the spread of misinformation on their platforms. However, they would prefer not to be the arbiters of truth as the truth can be subjective or otherwise hard to determine. Instead, they would prefer that social media users themselves show better discernment when deciding which information to share. Here we show that allowing people to share only those social media posts that they have indicated are true significantly improves sharing discernment, as measured by the difference in the probability of sharing true information versus the probability of sharing false information. Because it doesn’t require social media companies to be the arbiters of truth, this self-censorship intervention can be employed in situations where social media companies suspect that individuals are propagating misinformation but are not sufficiently confident in their suspicions to directly censor the individuals involved. As such, self-censorship can usefully supplement externally imposed (i.e. traditional) censorship in reducing the propagation of false information on social media platforms.


Self-Censorship Appears to be an Effective Way of Reducing the Spread of Misinformation on Social Media

Virtual Bargaining and the Moral Mind

Moral learners face an underdetermination problem – the rules they are taught cannot account for all novel cases. One way learners solve this problem is through closure principles, through which a learner assumes that anything that isn’t explicitly forbidden is permitted, and vice versa. The current work aims to explore whether closure rules are used when reasoning about obligations. Building on previous work, we ask 1) whether adults and children use obligation rules in a similar manner to rules about permissibility and impermissibility, and 2) how early in life these inference abilities emerge. Across two studies, we explore inferences about obligation from both adults (N = 120, Mage = 33.73 years) and children (N = 103, Mage = 5.52 years). We found that while both adults and children rationally learn closure principles consistent with deontic logic, children and adults make opposite inferences about novel cases when provided with obligation rules.


Underdetermination and Obligation Rules: Adult and Children’s use of Closure Principles in Moral Learning

Educational technologies with AI are designed to personalize students’ learning, but also to alleviate teacher’s workload. However, acceptability of such technologies among teachers may be impacted by factors such as fear of replacement or ethical concerns. The purpose of the current study is to investigate attitudes of teachers towards educational tools with or without AI. The main hypothesis is that technologies with AI would be more negatively judged than technologies without AI, and thus intention to use would be weaker for technologies with AI. Results show that teachers seem to accurately perceive the potential benefit of AI technologies for reduction of workload, without feeling threatened by a replacement in the future. Ethical concerns are higher for AI technologies, but intention to use is pretty much similar. Differences between primary and secondary teachers are discussed.

Teachers’ attitudes towards AI: what is the difference with non-AI technologies?

Adults are more accurate at identifying own- than other-accented talkers (e.g., Stevenage, 2012). However, no study to date has examined when the Other Accent Effect (OAE) first emerges. Here, we use a voice line-up task to test 6- to 12-year-olds (N=59) on their ability to recognize own- (Canadian English) and other-accented (Japanese- or Mandarin-accented English) talkers. Own-accented talkers (M=0.77) were recognized better than other-accented talkers (M=0.61), p<.005. No effect of age was observed. This indicates the OAE is well-entrenched by 6 years of age. Interestingly, all children tended to evaluate other-accented talkers negatively, but these biases did not predict the strength of the OAE, suggesting that a factor other than social biases drives the effect (Yu et al., 2021). To further understand the underlying mechanisms of the OAE, we are currently testing children in the same line-up task on their ability to recognize Southern American and Australian English talkers.

The Emergence of the Other Accent Effect in Talker Recognition

Political polarization is driven by many factors, but the role of moral values as both a signal of political identity and a source of internal conflict is understudied. We report an agent-based computational model of polarization that fills this gap. Agents seek to differentiate in- and outgroup neighbors with a slight preference for the former. However, they must do so by inferring neighbors’ identities from visible but transient moral signals. Moreover, agents experience conflicts within their own values, and if difficult to resolve internally, can copy the values of their ingroup or disengage (i.e., act immorally). Results show that liberals form larger, more homogeneous clusters, are happier, and experience less moral conflict than conservatives. Conservatives experience more and higher levels of conflict, and morally disengage significantly more often than liberals.


Simulating Political Polarization as a Function of Uncertain Inference and Signaling of Moral Values

Punishment is a cost imposed on a target, in response to an undesirable action. Yet choosing to punish also reveals information about the authority’s own motives and values. 
We propose that observers jointly infer the wrongness of the action and the authority's motivations. Using hypothetical scenarios in unfamiliar societies, we experimentally manipulated observers' prior beliefs and measured human observers' inferences after observing punishment. These inferences were recapitulated in a formal model that inverts an intuitive causal model of authorities who make rational choices about punishment by weighing its costs and benefits (i.e. utilities). An essential component of this model, driving these inferences, is that legitimate authorities consider the utility of a proportional response to harmful actions, which depends on the balance between the wrongness of the act and the severity of the punishment.

What people learn from punishment: joint inference of wrongness and punisher's motivations from observation of punitive choices

In the submodular ranking (SR) problem, the input consists of a set of submodular functions defined on a ground set of elements. The goal is to order elements for all the functions to have value above a certain threshold as soon on average as possible, assuming we choose one element per time. The problem is flexible enough to capture various applications in machine learning, including decision trees. 

This paper considers the min-max version of SR where multiple instances share the ground set. With the view of each instance being associated with an agent, the min-max problem is to order the common elements to minimize the maximum objective of all agents---thus, finding a fair solution for all agents. We give approximation algorithms for this problem and demonstrate their effectiveness in the application of finding a decision tree for multiple agents. 

Min-Max Submodular Ranking for Multiple Agents

With artificial intelligence systems increasingly applied in consequential domains, researchers have begun to ask how AI systems ought to act in ethically charged situations where even humans lack consensus. In the Moral Machine project, researchers crowdsourced answers to "Trolley Problems" concerning autonomous vehicles. Subsequently, Noothigattu et al. (2018) proposed inferring linear functions that approximate each individual's preferences and aggregating these linear models by averaging parameters across the population. In this paper, we examine this averaging mechanism, focusing on fairness concerns and strategic effects. We investigate a simple setting where the population consists of two groups, the minority constitutes an α < 0.5 share of the population, and within-group preferences are homogeneous. Focusing on the fraction of contested cases where the minority group prevails, we make the following observations: (a) even when all parties report their preferences truthfully, the fraction of disputes where the minority prevails is less than proportionate in α; (b) the degree of sub-proportionality grows more severe as the level of disagreement between the groups increases; (c) when parties report preferences strategically, pure strategy equilibria do not always exist; and (d) whenever a pure strategy equilibrium exists, the majority group prevails 100% of the time. These findings raise concerns about stability and fairness of averaging as a mechanism for aggregating diverging voices. Finally, we discuss alternatives, including randomized dictatorship and median-based mechanisms.

Moral Machine or Tyranny of the Majority?

Face image synthesis has progressed beyond the point at which humans can effectively distinguish authentic faces from synthetically-generated ones. Recently developed synthetic face image detectors boast ``better-than-human'' discriminative ability, especially those guided by human perceptual intelligence during the model's training process. In this paper, we investigate whether these human-guided synthetic face detectors can assist non-expert human operators in the task of synthetic image detection when compared to models trained without human-guidance. We conducted a large-scale experiment with more than 1,560 subjects classifying whether an image shows an authentic or synthetically-generated face, and annotating regions supporting their decisions. In total, 56,015 annotations across 3,780 unique face images were collected. All subjects first examined samples without any AI support, followed by samples given (a) the AI's decision (``synthetic'' or ``authentic''), (b) class activation maps illustrating where the model deems salient for its decision, 
or (c) both the AI's decision and AI's saliency map. Synthetic faces were generated with six modern Generative Adversarial Networks. Interesting observations from this experiment include: (1) models trained with human-guidance, which are also more accurate in our experiments, offer better support to human examination of face images when compared to models trained traditionally using cross-entropy loss, (2) binary decisions presented to humans results in their better performance than when saliency maps are presented, (3) understanding the AI's accuracy helps humans to increase trust in a given model and thus increase their overall accuracy. This work demonstrates that although humans supported by machines achieve better-than-random accuracy of synthetic face detection, the approaches of supplying humans with AI support and of building trust are key factors determining high effectiveness of the human-AI tandem.

Underline digital video library

The REAL AI revolution: the benefits of Chatbot Augmented Instruction

Using Simulation Practices to Refine Clinical Skills for Hearing Screenings

Self-Censorship Appears to be an Effective Way of Reducing the Spread of Misinformation on Social Media

Virtual Bargaining and the Moral Mind

Underdetermination and Obligation Rules: Adult and Children’s use of Closure Principles in Moral Learning

Teachers’ attitudes towards AI: what is the difference with non-AI technologies?

The Emergence of the Other Accent Effect in Talker Recognition

Simulating Political Polarization as a Function of Uncertain Inference and Signaling of Moral Values

What people learn from punishment: joint inference of wrongness and punisher's motivations from observation of punitive choices

Min-Max Submodular Ranking for Multiple Agents

Moral Machine or Tyranny of the Majority?

The Value of AI Guidance in Human Examination of Synthetically-Generated Faces

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES