The Federal Communications Commission (FCC) oversees the regulation of Internet Protocol Captioned Telephone Service (IP CTS), an assistive technology that enables people with hearing loss to use their own voice to speak during a call, and then to read the other party’s responses via captions on a home telephone device or app. IP CTS is commonly referred to as captioned telephone service.
Certain voices and noises create difficulty for Artificial Intelligence (AI), including Automated Speech Recognition (ASR). This video compares real-time human captioning with ASR captions.
Captioned telephone service is funded through a small fee on consumer phone bills, making the program available to Americans with hearing loss. As use of these services continues to grow, however, the FCC is concerned about the increasing costs of the program and is considering changes to reduce them.
The Americans with Disabilities Act (ADA) requires that people with hearing loss have access to communications technology that is “functionally equivalent” to what people without hearing loss use. Currently, captioned telephone services use a communications assistant who is either a stenographer or a professional who uses Automated Speech Recognition (ASR) to create captions of what a caller says. Today, the human element ensures accuracy of the captions. The FCC is considering certifying vendors that offer ASR-only solutions for captioned telephone service. This would take the human element out of the equation, similar to Apple’s Siri or automated closed captioning on YouTube. But, as of now, the FCC does not have the quality standards in place to ensure functional equivalency of such ASR-only captioning. While ASR is well on its way to becoming a viable option, it’s simply not yet accurate enough for use on its own for captioned telephone service.
Without the human transcribers to ensure accuracy, the quality of captioned telephone service would be greatly diminished. ASR technology cannot account for issues like background noise, accents or speech disabilities, which can lead to significant inaccuracies in captioning. These inaccuracies pose safety issues. Consider a situation in which someone is speaking to a doctor or emergency responder and receives inaccurate instructions via captioning. Access to accurate captions is crucial for all who use captioned telephone services.