About the Authors:  Dr. Hope Gerlach-Houck is an assistant professor at Western Michigan University. Her research explores communication stigma and ableism and focuses on making a meaningful impact on the lives of people who stutter. She serves as the Publicity and Marketing Lead for HeardAI. Caryn Herring is a person who stutters, speech-language pathologist, and a doctoral candidate at Michigan State University. Her research interests include the process of desensitization and reducing the adverse cognitive and affective components of stuttering. She serves as the Accessibility and Fairness Lead for HeardAI. Dr. Ann Marie Ryan is a professor of Organizational Psychology at Michigan State University. Her research explores fairness in hiring processes and fosters workplace inclusion. She serves as the Policy and Advocacy Lead for Heard AI and brings insights on the use of automatic speech recognition in work contexts as well as on workplace accommodations for those with speech differences. Dr. Scott Yaruss is a professor of Communicative Sciences and Disorders at Michigan State University. His NIH- and NSF-funded research focuses on reducing the adverse impact that people who stutter experience in their lives. He serves as the Stuttering Data Lead for HeardAI. Dr. Nihar Mahapatra is an associate professor at Michigan State University and the principal investigator for HeardAI. His research focuses on AI accessibility for disability inclusion and environmental sustainability, with a special emphasis on voice AI for people with speech differences and smart species ID for marine conservation. Nihar has also explored AI-driven recruitment to shape the future of work and modeled protein-ligand interactions for early-stage drug discovery. Dena Mujtaba is a doctoral student at Michigan State University studying Electrical and Computer Engineering. Her research interests include AI fairness, automatic speech recognition, and natural language processing. She works as a research assistant for HeardAI. Megan Arney is a speech-language pathologist and doctoral student at Michigan State University. Her research interests include stuttering, specifically the impact that this has on a person’s life. She works as a research assistant for HeardAI. Jia Bin is a doctoral student at Michigan State University. Her own journey of being a person who stutters, and an inter-cultural communicator motivates her to pursue a doctorate to help the people who stutter and bring people from various cultural backgrounds together. She works as a research assistant for HeardAI. Dr. Chelsea Johnson is the project manager for HeardAI. She is passionate about increasing accessibility for people who stutter and strives to create an inclusive world where everyone has equal opportunities and access to the resources they need.

Listening is critical for meaningful interactions, building connections, and meeting personal needs. Good listeners can make all the difference in the day-to-day lives of people who stutter. Unfortunately, society is not always aware or considerate of how to be good listeners, particularly when speaking with someone who stutters. Poor listening can take many forms, including interrupting, finishing sentences, not attending to what someone is saying, and more. When asked to imagine a “bad listener,” most people would visualize a person, but technology can also demonstrate poor listening skills that result in inconvenience, inaccessibility, and inequity for people who stutter. In the context of voice technology, consequences of “bad listening” can include device activation failures, speech transcription inaccuracies, errors in speech recognition, and more.

Voice AI technology is integrated into billions of devices, revolutionizing life and work for many. Yet, for the more than 80 million people worldwide who stutter, along with over 180 million additional people with other communication differences, these technologies often fall short. Unfortunately, existing automatic speech recognition (ASR) systems, such as those found in our phones (e.g., Siri), phone trees (e.g., “Say in a few words why you are calling”), and automated processing of job interview responses, cannot consistently understand stuttered speech. In the workplace, this can result in unfair screening of job applicants based on incorrect inferences drawn from poorly transcribed interviews. In other contexts, speech recognition failure can result in incorrect purchases or charges, wrong directions, and increased frustration for people who stutter. Thus, the power and benefits of improved listening are relevant not only for human listeners; they are also important for digital listeners. 

We are a team of interdisciplinary researchers, people who stutter, speech therapists, engineers, and psychologists who are dedicated to solving this problem through our new initiative, HeardAI. Funded by the National Science Foundation, HeardAI is a technology and advocacy project with the mission of making voice AI more broadly usable and equitable. Our deep connection with the stuttering community is a key differentiator of our work. People who stutter serve as partners, advisors, and participants in design decisions. They contribute data for training our AI and provide feedback on all aspects of our project, ensuring that our solutions are effective and truly inclusive. Our leadership team and primary research assistants include people who stutter and allies, thereby reinforcing our commitment to the community and ensuring that their perspectives are represented at the highest level. Learn more about our vision and approach here. 

In order to create accessible technology, stuttered voices need to be represented in the datasets that voice AI systems are trained on. If you are an adult who stutters (18+) who is proficient in English and currently lives in the United States, you are invited to participate in our study that involves collecting speech samples from people who stutter. Please contact us at the email below for more information. If you are interested but not from the United States, please know that we plan to open this study to other countries soon. 

In addition to completing our study, we would love to learn more about your personal experiences with voice technology. Using the discussion threads for this post, please tell us about your experiences with voice AI and let’s get a conversation going. Consider the following:

• Have you had any memorable experiences with Voice AI, either positive or challenging? 
• How does Voice AI’s accessibility, or lack thereof, influence your daily life or work? 
• How does the theme of “The power of listening” resonate with your experiences with Voice AI? 
• What features or improvements would you like to see in voice AI to better support your needs? 

If you have questions or would like to get in contact with us, email us at HeardAI.team@heardai.ai