A recent Fortune magazine report revealed that artificial intelligence-powered search engines are inaccurate more than 60% of the time, a deficiency that extends to AI-generated captions and poses significant accessibility challenges for millions.
The findings highlight a growing concern as organizations increasingly rely on automated captioning solutions for online videos, webinars, and digital learning platforms. Gay Cordova, founder of CCTubes, a human-created captioning and accessibility service, emphasizes the critical importance of accuracy and quality in captioning. “If you want your message to get across, you need those two things,” she said.
Cordova, a former court reporter who transitioned to captioning in 2005, notes a dramatic shift in the demand for captioning over the past two decades. Initially driven by broadcast regulations requiring 75% of content to be captioned, the proliferation of streaming services and online content has created an unprecedented need across industries. “Captions are no longer just for the few; they’re now a universal accessibility issue,” Cordova explained.
The need extends beyond the deaf and hard-of-hearing community. The World Health Organization estimates that over one billion young people globally are at risk of hearing loss due to unsafe listening practices. Partial hearing loss, coupled with poor audio quality and diverse accents, can develop comprehension difficult for a growing population. Cordova has observed a rise in caption usage even among individuals without hearing impairments. “If you’re in a noisy environment or the audio isn’t clear, you turn them on,” she said.
While acknowledging the appeal of AI-driven automation for its speed and scalability, Cordova cautions against compromising accuracy. She maintains a 98% accuracy standard for her human captioners, a benchmark she believes AI frequently fails to meet. “Imagine relying completely on captions and missing so much of the information. Missing a modifier can alter instructions, and absent punctuation can obscure tone and structure altogether,” she stated.
The challenges are particularly acute in live environments and when dealing with regional accents or specialized terminology. In corporate training, compliance briefings, and academic lectures, inaccuracies can compromise effectiveness and inclusivity. “If someone hard of hearing reads the captions that persistently leave out key words, that can be detrimental to their performance. It’s not okay to make anyone feel like they don’t know what’s going on,” Cordova shared. She recounted the gratitude expressed by deaf students and professionals who benefit from accurate, human-generated captions, stating, “They are so thankful. Every class, they say, ‘Thank you for coming and helping me.’ They have been forgotten in many spaces.”
Cordova argues that human captioners offer a level of interpretive nuance that AI currently lacks. They can understand context, identify speakers, apply punctuation strategically, and resolve ambiguity. This interpretive layer, she believes, transforms captions into a meaningful communication channel. Beyond accessibility, Cordova positions quality captioning as a strategic investment for content creators, enhancing search engine optimization, viewer engagement, and global reach through translation workflows, ultimately reinforcing brand credibility. “When you caption your content, it’s done once, and it can be replayed forever,” she explained. “It may cost more upfront, but you reap the rewards.”
Cordova advocates for a hybrid model that combines technological advancements with human expertise to ensure accessibility and accuracy as content ecosystems continue to expand. She views captioning as a fundamental responsibility, ensuring that communication reaches its intended audience with integrity, and that comprehension is prioritized alongside efficiency.
