The answer is ‘No,’ by the way… and here’s why…
For a few years now, predictions have been made on the growth of AI or AI captioning services taking over the internet, with many suggesting that by 2026, 90% of online content will be generated by AI.
For a long time now, large companies have been progressively turning to Artificial Intelligence in order to push down costs and increase revenue. Smaller companies for a long time, bucked and resisted that trend, keeping captioning, translation and even voiceover services human. But with global companies offering lower and lower costs to use AI, sadly, many small businesses have also turned to the use of AI in order to try and keep up with competitive pricing.
At Capital Captions, the amount of work received from video production companies and indeed, other clients hasn’t reduced, but it has definitely changed in scope.
Often times, we find ourselves recreating captions which have already been created by AI, but just aren’t up-to-scratch, and need a human to fully QC and correct them. The vast majority of the time, these projects are faster completed by ‘going back to the drawing board’, and totally starting again.
Capital Captions services have also shifted from closed captioning and translation, to video accessibility on the whole. We accept from clients working on broadcast projects, film festivals, or video series and films that need to meet accessibility requirements. In short, ‘the complicated stuff.’
We provide professional editing and accessibility services, working with subtitling, closed captioning, translation, BSL interpreting and audio description voiceovers. Our focus end and goal is always on true video accessibility.
Where captioning work online often could involve subtitling YouTube videos and social media content, many are turning to AI captioning for these small projects – some successfully, others, not so much.
Short online videos often feature subtitles that are AI generated and burned in to be more like graphics, with one word of speech flashing up at a time. In a world of video shorts to match equally short attention spans, these captions are made to look visually exciting but are often hard to read, and flick away so fast that they actually detract from what is being said.
In most cases, they don’t really aid accessibility at all. For anyone that truly cares about using subtitles for accessibility, to aid listening and understanding, this new AI-driven captioning style is incredibly frustrating.
As an aside, the rise of video shorts as a way for companies and influencers to gain audiences ironically has the opposite impact overall. AI generated shorts are often posted as means of ‘advertisement’ to entice viewers to a larger channel, website or brand.
But the huge number of video shorts online has unintended consequences. Viewers’ attention spans are becoming shorter and rather than clicking away to view a longer video, or indeed, browse any product, website or service, viewers are instead choosing to scroll endlessly from short to short. Ultimately, lazy AI video promotion end up going hand-in-hand with lazy AI captions.
When it comes to closed captioning and subtitling for longer online videos, conferences and the like, AI comes with a whole new range of problems. Whilst captioning on video shorts are almost entirely useless in terms of video accessibility, long form videos with AI captions can be nonsensical, frustrating and at times, outright hilariously bad.
Voice recognition consistently makes mistakes in deciphering accents, identifying sentences when over-speaking happens in conversation and coping at all with any issues in sound quality. Similarly, AI often makes incorrect grammatical choices and typically types everything absolutely verbatim, including every single ‘ah,’ ‘um,’ ‘you know,’ ‘kind of,’ ‘sort of’, ‘so,’ ‘well’… You get the idea!
Whilst guidance on writing closed captions often recommends using verbatim as much as possible, professional captioners will filter out excessive use of these words, not only because it makes subtitles easier to read, but also because including every single word often means reading speeds increasing to such an amount that there’s not even time to read every word anyway.
The usual recommendation with captioning long form videos in AI is to have the subtitles auto-generated and then have a human watch them back, making amendments and corrections as they go. Often times, there are so many mistakes that this just doesn’t happen. Companies and individuals carrying out their own AI subtitle checks will get bored and give up early on, and those who care enough not to quit, often wind up having those captions QC’d or recreated entirely by a professional captioning company.
Whilst we strongly advise against AI being used for captioning, for some videos where speech is absolutely clear, straightforward and with no accents, we accept AI can sometimes do a reasonable job. When it comes to translation, however, it’s a whole new story.
AI translation has a very, very long way to go. Translation requires a much fuller understanding of what is being said than transcription does. Languages aren’t translated word for word. Localisation needs to happen. Sayings need to be made sense of in a way that is relevant to the target language. Jargon needs to be described and explained, whilst product and company names need to be left in the source language in order to make any sense at all. AI cannot make these choices. Not even close.
In an ever-increasingly automated world, at Capital Captions, we’re holding strong. We believed in humans before, we believe in them now and we plan to go on believing in them. With our prices remaining stable (almost exactly the same rates as they were 8 years ago), we still only work with professional subtitlers, translators, transcriptionists and voice artists, and aim to keep our prices competitive.
The way we see it, AI captioning services have ultimately been brought in for use as a cost saving exercise. People use AI captioning because it’s cheap, but even if that’s the case now, as time goes on, those prices will increase. Captioning companies such as ours, have struggled in recent years to keep up with lowering rates when competing against incredibly cheap AI alternatives. But inflation is real, and tech companies get greedy.
As AI translation and captioning prices increase, those irritating (but cheap) AI mistakes that everybody puts up with because they were getting cheap services, aren’t going to feel like they’re worth the hassle and the tide will turn back. That’s our theory anyway, and regardless of what happens, Capital Captions human services are here to stay.
So there you have it, are AI captioning services better than human captiopns? The definitive answer is NO. If you would like to know more about our human closed captioning services or transaltion services or you know what you would like then click the button below for your quote today.
Captions created for deaf and hard of hearing
They include spoken word, identification of speakers and descriptions of effects
Accessibility for the blind and visually impaired
visually descriptive events as they appear in the video and on-screen.
Subtitle translation for different languages
Working with best transcriptionist and linguists for best possible outcome
March 12, 2025
November 11, 2024
July 22, 2024
March 12, 2025
November 11, 2024
Sunday 8am-5pm
Capital Captions specialise in high quality subtitling, closed captioning, video transcription, translation, and voiceover services. Our flexible approach to working with video content means we can effectively provide a one-stop-shop for clients wanting to turn their audio into text.
March 12, 2025
November 11, 2024
July 22, 2024
Cookie | Duration | Description |
---|---|---|
cookielawinfo-checkbox-analytics | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Analytics". |
cookielawinfo-checkbox-functional | 11 months | The cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional". |
cookielawinfo-checkbox-necessary | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary". |
cookielawinfo-checkbox-others | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Other. |
cookielawinfo-checkbox-performance | 11 months | This cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Performance". |
viewed_cookie_policy | 11 months | The cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies. It does not store any personal data. |
Comments are closed.