What will become of the captioning industry if the marketing methods by Automatic Speech Recogntion (a/k/a Artificial Intelligence captioning) companies convince clients to switch from human captioning?
Already I see a new model taking shape. A salesman I once knew was told by his boss to “Sell the sizzle!” I see a lot of “sizzle selling” out there, marketing “new and improved” ASR/AI captioning – “at a fraction of the cost of human captioning.” S-S-S-SIZZLE!
What responsible, budget-minded company or agency wouldn’t want to save money?!
The catch is that the ASR/AI products are demonstrated in controlled environments, with people individually speaking into their own dedicated microphones, say in a Zoom (or similar) platform, and identified by whatever name (or initials) they sign in with. Who is saying what? A user of captions needs to know quickly who is speaking, else is hampered from following information equally.
Further, if ASR/AI's exteensive bank of sounds doesn’t produce a match due to accents, overlapping speech, barking dogs, or idiomatic references, it just drops the words, with no attempt to paraphrase what was said or put in any meaningful clues like [CROSS-TALK] or [OFF-MIC], which is what human captioners do.
A human captioner will interrupt politely to ask for a repetition or clarification so that users of the captions -- usually persons who are deaf, hard of hearing, or learning-disabled, or those who natively speak other tongues – have equal access to communication, mandated under the Americans with Disabilities Act.
The final irritation for me is when I learn that an AI companies use human captioners to caption their meetings or demonstrations of "their product" to a client. The client can be so enamored of the "sizzle" that the ASR/AI company “wins” the business or contract on cost and the dishonest demo.
My email box this week had two requests from ASR/AI captioning companies to send them my rates and certifications for realtime captioning work. Hmm, why do they need these if their product is so great? Could it be that the AI captioning companies need us for the BAIT and SWITCH? Sell ‘em the sizzle, then upsell human captioning at a premium rate to the client? That is already the business model for several AI captioning companies. “For an added fee, we can give you the Premium level” – which is, of course, us human captioners.
Many clients who have tried "the sizzle" of ASR/AI come back to human captioners. The threat of the sizzle-sale is certainly forcing human captioners to write at their very best. ASR/AI may be great for "free" notes of a meeting to someone who heard what was said, but not when equal access is needed and, by law, mandated.
Gayl Hardeman, co-owner of Hardeman Realtime, Inc.
Comments