Can a Deaf Candidate Pass Your AI Voice Screen? What Indian Disability Law Already Requires

Every AI voice screening vendor pitching campus hiring in India, HireQwik included, has to answer a question that rarely comes up in the sales deck: what happens to the candidate who can’t take a 15-minute voice call because they’re deaf, hard of hearing, or have a speech disability. The Rights of Persons with Disabilities Act, 2016 already requires employers to not discriminate in recruitment on this basis, and it requires establishments to make their processes accessible, not someday, now. Most AI screening rollouts in India haven’t sat down and answered the question directly.

The law is older than the technology gap

The RPwD Act isn’t a new or theoretical compliance risk. It’s a 2016 statute that applies to recruitment specifically, and it sits alongside a broader digital-accessibility expectation: establishments are expected to make digital services, including recruitment platforms, accessible by following recognized standards like WCAG. What’s changed is the delivery mechanism for that first-round screen. A phone-screen recruiter can (and does) accommodate a candidate who signs, types, or needs an interpreter, often without anyone formalizing a policy for it. A structured, voice-only AI interview has no such built-in flexibility unless someone designs one in.

As the International Bar Association has pointed out, India doesn’t yet have anything like the EU AI Act’s high-risk classification for hiring AI, and there’s no dedicated legal architecture forcing bias audits or accessibility reviews of automated screening tools specifically. That gap cuts both ways: it means less near-term regulatory pressure than a GCC serving EU clients might expect from the EU AI Act side, but it also means the existing RPwD obligation doesn’t come with a ready-made technical checklist. HR teams adopting voice AI are, in practice, the ones deciding what “accessible” means for their own process.

Where a voice-only screen genuinely breaks

Be specific about where the format fails, because vague reassurance is worse than no answer. A candidate who is deaf or hard of hearing cannot take a live voice call at all. A candidate with a significant speech disability may be scored on fluency and pace metrics that were never designed to separate “disability” from “poor communication fit,” and an AI model has no way to tell the difference unless it’s explicitly built to. Neither of these is a minor edge case decided by an accent model tweak; they require an actual alternate pathway (text-based screening, an interpreter-assisted format, or a human fallback) offered before the candidate hits a wall, not after a rejection triggers a complaint.

This is not a problem unique to HireQwik. It’s a problem for every vendor selling “AI screens every candidate the same way” as the pitch. Consistency is the feature, right up until the format itself excludes someone the law says you can’t exclude.

What we tell HR teams evaluating this

We don’t have a polished answer that makes this disappear, and pretending otherwise would be the kind of overclaiming that erodes trust faster than any single compliance gap. What we do tell HR teams evaluating voice AI screening, us or anyone else, is to ask three concrete questions before rollout: does the platform have a documented alternate-format pathway for candidates who disclose a hearing or speech disability; who reviews a rejection that might correlate with a disability-linked speech pattern, and how quickly; and is that pathway visible to candidates at the point they apply, not buried in a footer link. If a vendor can’t answer those three questions with specifics, the AI-screens-everyone-consistently pitch is incomplete, not wrong.

For a company running a 3,000-candidate campus drive, the honest framing is that automated voice screening handles the overwhelming majority of that pool well and handles a small, legally-protected minority of it badly if nobody designed for them. Getting the first part right doesn’t excuse skipping the second. Building the alternate pathway is the harder, less glamorous slice of the job, and it’s the part that keeps a fast, cheap screening process from becoming a liability instead of an advantage.

If you’re rolling out AI voice screening this hiring season and want a straight answer on where the accessibility gaps sit in your specific process, talk to us before you launch — not after the first complaint.