CaptionCall's automatic captions beat assisted ones in FCC lab tests

In the FCC's National Test Lab, CaptionCall's automatic captions had median error rates of 2.9 to 14.5 across five scenarios. Human-assisted providers had 8.9 to 19.5 on average.

  • Not rated a person

    The FCC approved the automatic mode on test results before it captioned users' calls, so no outcome for users is reported.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Intended benefit.
Evidence

foundstudiedreportedclaimed

An independent regulatory document supports the benefit.

Compared with

Median error rates of CA-assisted providers on the same five call scenarios. Measured.

Without the system
8.9 to 19.5 on average
With it
2.9 to 14.5
The system’s part
The National Test Lab measured the automatic captions against those of human-assisted providers on the same call scenarios, which isolates the platform's output. The assisted captions also start from speech recognition, which the assistants feed by revoicing. The vendor is unnamed, and the FCC does not show that the platform learns from data.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Not yet checked.

Download data Cite Suggest a correction

Benefit type
Improved Task Accuracy
System
CaptionCall ASR-only (fully automatic) Internet Protocol Captioned Telephone Service (IP CTS)
Made by
CaptionCall, LLC (with Sorenson Communications, LLC), on a speech recognition platform from an unnamed vendor
Used by
CaptionCall, LLC
Where
United States
Stage
Tested in a laboratory
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Documenting/Recording Information
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Regulatory document, by authors with no tie to the system.

    Federal Communications Commission, Consumer and Governmental Affairs Bureau. Order, DA 21-1080, CG Docket No. 03-123 (conditional certification of CaptionCall, LLC to provide IP CTS on a fully automatic basis). Adopted and released 31 August 2021.

    The National Test Lab’s own testing provides further evidence that, in terms of accuracy, CaptionCall’s ASR platform can outperform CA-assisted IP CTS. In repeated tests of five call scenarios, CaptionCall’s ASR platform achieved median Word Error Rates ranging from 2.9 to 14.5, while CA-assisted providers’ median Word Error Rates ranged from 8.9 to 19.5 on average.
    CaptionCall’s median per-word caption delays for various call scenarios ranged from 3.3 to 3.8 seconds, while CA-assisted providers’ median caption delays were significantly longer, averaging from 5.2 to 17.8 seconds.

Details

AI system

Workflow

CaptionCall chose its recognition platform after comparing leading vendors on accuracy and formatting. Calls go to the automatic platform when no communications assistant is available.

Follow-up

Conditions for replication

Vendors may not keep users' data or the text of captions. The automatic captions must still meet the FCC's minimum standards once in use.

Entry details

Sector

Telecommunications relay services / accessibility for people with hearing loss.

Purpose

CaptionCall built its fully automatic mode to give people with hearing loss captions of their phone calls when no communications assistant is available.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). CaptionCall's automatic captions beat assisted ones in FCC lab tests (AIBD-0824). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0824.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions