A Phenomenal ASL–English Video Podcast on Risk, Responsibility, and Reality
In this timely episode, podcast host Eddie Arrieta facilitates a substantive and informed conversation with Jeff Shaul and Molly Glass from CoSET (Coalition for Sign Language Equity in Technology) alongside Katharine Allen and Mila Golovine from the Interpreting SAFE AI Task Force.
A Holistic Framework for AI Interpreting
The four guests respond to Eddie’s questions about the essential paradigm shift described in the AI Interpreting Solutions Evaluation Toolkit (2025). Part A of the Toolkit takes a holistic approach to risk governance for organizations thinking about purchasing of this revolutionary new sociotechnical communication system. Interpreting is treated as a reciprocal communication process between primary communicators involving:
- Technology – AI capabilities, limitations, and failure modes
- Human participants – primary communicators of two or more languages, and professional interpreters
- Context – legal, medical, educational, and business environments
- Organizational decision-making – governance, accountability, and liability
The first element is technical, the remaining three are social aspects of language access: these are interactive, ethical, and fluid. This novel approach of Part A rejects the idea that AI interpreting can be evaluated only at the level of an AI’s language model performance.
Why This Matters Now
Companies are rapidly bringing automated AI “solutions” to the market. To learn more about the risk picture and how to assess automated artificial intelligence interpreting systems, watch this episode of Localization Today, a contemporary podcast covering technological developments for delivering language access.
By watching the podcast, you will learn about recommended consumer controls, and why buyers of AI interpreting services need to use standardized checklists for bottom-line safety. Technology developers are encouraged to collaborate with primary communicators and professional human interpreters to anticipate potential failure modes and accelerate the development of S.A.F.E. AI interpreting software—systems that are Safe, Accountable, Fair, and Ethical.
Watch and Engage
Watch the video-podcast “ Inside the AI Interpreting Solutions Evaluation Toolkit “ and see what you think!
WANT MORE?
To receive $5 off the next upcoming workshop on Part B of the AI Interpreting Solutions Evaluation Toolkit, share this blog post on social media using #DeafSafeAI. When the workshop is announced, you’ll be able to include the URL to your social media post.