Captions in Crisis: The Uncomfortable Truth Behind British Broadcasting's Accessibility Gap

When Ofcom publishes its annual accessibility report, the headline figures tend to read encouragingly. The major public service broadcasters meet or exceed their subtitle quotas. Compliance boxes are ticked. Press releases are issued. Yet for the estimated twelve million people in the United Kingdom who are deaf, hard of hearing, or reliant on captions for any number of reasons, the numbers on a regulator's spreadsheet bear little resemblance to the nightly frustration of watching garbled text scroll across a programme they paid a licence fee or subscription to enjoy.

The gap between regulatory compliance and genuine viewer experience has widened considerably over the past five years — and the causes are neither simple nor confined to a single broadcaster.

The Letter of the Law, and Little Else

Ofcom's Broadcasting Code requires that qualifying channels meet specific targets for the proportion of output that carries subtitles, audio description, and signing. For the main PSB services, the subtitle target sits at one hundred per cent of applicable programming. On the surface, this sounds like a solved problem.

It is not. The regulations govern quantity, not quality. A broadcaster can satisfy its legal obligation by supplying subtitles that are riddled with errors, poorly timed, or formatted in a manner that makes them genuinely difficult to read — and face no formal sanction for doing so. This structural loophole has allowed a two-tier system to develop quietly, one in which certain broadcasters invest meaningfully in caption quality while others treat accessibility as an administrative exercise.

"The compliance framework was designed at a time when subtitling was an expensive, manual process," notes one production executive with experience across both terrestrial and streaming delivery. "The rules haven't kept pace with what technology can now achieve, or indeed with what technology can get catastrophically wrong."

The AI Problem

The arrival of automated speech recognition as a subtitle-generation tool has been, depending on one's perspective, either a democratising breakthrough or a quality catastrophe. For smaller channels and streaming platforms with vast back catalogues, AI-driven captioning has made it economically viable to subtitle content that would otherwise have gone without. That is a genuine gain.

However, the same technology, when deployed carelessly or without adequate human review, produces results that range from mildly confusing to entirely unintelligible. Viewers and accessibility advocates have documented instances of proper nouns mangled beyond recognition, regional accents generating wildly inaccurate transcriptions, and technical or medical terminology replaced with phonetically similar but meaningless substitutes. In drama, where dialogue carries narrative weight, such errors do not merely inconvenience — they exclude.

The Royal National Institute for Deaf People has repeatedly raised concerns about the uneven quality of AI-generated subtitles across British platforms. The organisation's position is unambiguous: automated captioning without meaningful human oversight is not accessibility provision in any meaningful sense.

Where the Standard Is Being Set

The picture is not uniformly bleak. The BBC, whose obligations under its charter extend beyond Ofcom's minimum requirements, has invested substantially in its subtitle workflows, including dedicated quality-assurance processes and a long-standing commitment to live subtitle accuracy that has improved markedly over the past decade. Channel 4, which has positioned accessibility as a component of its broader social remit, has similarly demonstrated that quality and compliance need not be in tension.

Perhaps the most instructive case study from recent years is Channel 4's approach to its drama commissions, where subtitle style guides — governing font size, positioning, speaker identification, and reading speed — are treated as production documents rather than afterthoughts. The result is a consistency of caption quality that deaf viewers and advocates have praised publicly.

Some independent production companies have also begun embedding accessibility considerations earlier in the production pipeline, ensuring that captioning requirements are discussed at the scripting and post-production planning stages rather than bolted on at delivery. This shift in workflow, modest in cost but significant in outcome, is increasingly being championed by the Production Guild and other industry bodies.

The Streaming Blind Spot

If the PSBs represent the more regulated end of the spectrum, the streaming sector presents a considerably murkier picture. Subscription video-on-demand platforms operating in the United Kingdom are subject to Ofcom's on-demand accessibility requirements, but enforcement has historically been light-touch, and the pace of catalogue growth has consistently outstripped the provision of quality captions.

Viewers using services from both domestic and international providers have reported significant inconsistencies: flagship original productions captioned to a high standard sitting alongside acquired content where subtitles are plainly machine-generated and unreviewed. For a deaf viewer attempting to navigate a platform's library, the experience can be one of perpetual uncertainty — never knowing, until the programme begins, whether the captions will be adequate.

Ofcom's updated on-demand accessibility guidance, published in recent years, has tightened expectations somewhat, but the absence of prescriptive quality standards — as opposed to provision targets — means the fundamental problem persists.

The Industry Pressure Points

Speak to producers, post-production houses, and broadcasters about subtitle quality, and a consistent set of constraints emerges. Budget is the most frequently cited. Subtitling, when done properly by experienced professionals, carries a cost that squeezes already pressured production budgets. The temptation to substitute human expertise with automated tools — or to accept a lower standard of output from a cheaper supplier — is real and understandable, even if the consequences fall upon the viewer.

Deadlines compound the problem. In a content environment defined by rapid turnaround, the time required for thorough subtitle review and correction is often the first casualty of a schedule under pressure.

There is also, frankly, a cultural dimension. Accessibility has not always been treated as a creative priority within British television. The industry has made progress, but the perception of subtitling as a technical add-on rather than an integral element of the viewing experience persists in certain quarters.

A Standard Worth Having

What the industry requires, and what advocates have long argued for, is a quality framework to sit alongside the existing quantity targets — one with teeth. Prescriptive standards covering accuracy rates, synchronisation tolerances, formatting requirements, and review processes, applied consistently across broadcast and on-demand services, would create the conditions in which genuine improvement becomes an industry norm rather than a voluntary distinction.

The technology to achieve high-quality subtitling at scale exists. The regulatory architecture to mandate it does not yet, fully, exist. Closing that gap is not merely a matter of legal tidiness. For millions of British viewers, it is the difference between full participation in the nation's broadcasting culture and perpetual exclusion from it.