Voice – Code of Practice per KMA 2019

Last Updated on July 27, 2026 by Craig Allen Keefner

Below is a comprehensive summary of the draft Code of Practice – Tactility, Voice Recognition and Speech Command, with emphasis on the voice technology provisions and the regulatory implications for public self-service kiosks. The “meat” of this is the work of Peter Jarvis of Storm Keypads (retired).

Summary – Voice Recognition, Speech Command and Accessibility Regulations for Public Kiosks

The draft Code of Practice (CoP) proposes minimum accessibility standards for public ICT kiosks, including ticketing machines, healthcare kiosks, government terminals, self-order systems, payment kiosks and other self-service devices. The objective is to ensure equal access for users with visual, hearing, mobility, dexterity and cognitive disabilities while allowing vendors to innovate beyond the baseline requirements. The document emphasizes that compliance with the Code does not automatically satisfy legal obligations under ADA, Section 508 or other national laws, but instead establishes practical industry guidance.

Regulatory Background

The document positions itself as a bridge between today’s accessibility regulations and the next generation of AI-enabled kiosk interfaces.

Current regulatory references include:

• ADA 2010 Standards for Accessible Design
• Section 508 ICT standards
• DOT Air Carrier Access Act (ACAA)
• ICT Refresh Final Rule

Rather than replacing these regulations, the Code extends them into areas where formal standards are still developing—particularly speech interfaces and AI-driven accessibility.

Accessible Kiosk Population Requirements

The document recommends that accessibility should become the norm rather than the exception.

Its primary recommendation is that:

• At least 25% of kiosks serving a common purpose should be fully accessible.
• Individual standalone kiosks should always be accessible.
• New deployments should continue installing only accessible kiosks until the 25% threshold is achieved.

The recommendation follows the precedent established by the Air Carrier Access Act for airport kiosks and suggests expanding the same philosophy across all public self-service sectors.

Multi-Modal Accessibility Philosophy

A central concept throughout the draft is that no single interface works for every user.

Instead, future kiosks should combine:

• visual interfaces
• tactile controls
• audible guidance
• speech recognition
• voice command

Users should be free to select whichever combination best meets their individual needs.

This multi-modal approach is presented as the long-term direction for accessibility.

Audible User Interfaces

The proposed audible interface goes well beyond current ADA requirements.

The kiosk should provide:

• spoken instructions
• spoken navigation
• spoken descriptions
• spoken confirmation
• spoken transaction guidance

Audio should become available immediately when:

• a handset is lifted, or
• a headset or hearing device is connected.

The kiosk should then continuously repeat a welcome message until the user begins interacting, giving blind users unlimited time to prepare before navigation starts.

Audio Playback Requirements

The draft specifies several operational requirements.

Users must be able to:

• pause narration
• repeat previous instructions
• replay menu choices
• abandon a transaction
• restart from the welcome message

These recommendations recognize that blind users often require additional time to process spoken information.

Volume Control Standards

For kiosks located in noisy environments, the Code recommends:

• adjustable volume
• minimum 65 dB output
• up to 20 dB above ambient noise
• automatic reset to default volume after each session

The document also requires tactile volume controls that remain usable throughout the user session.

Voice Recognition and Speech Command

The most forward-looking section concerns Voice Recognition (VR) and Speech Command (SC).

The authors acknowledge that these technologies are:

• not currently mandated
• likely to become regulated
• rapidly advancing because of AI
• capable of dramatically improving accessibility

The draft argues that speech interfaces could become one of the biggest advances in independent living by allowing users to navigate kiosks naturally without relying entirely on touchscreens or keypads.

AI Changes the Accessibility Landscape

The Code recognizes two major technology trends occurring simultaneously:

1. AI-powered conversational interfaces
2. Voice biometrics for identity verification

Together these developments are expected to transform public kiosks.

Instead of forcing every user through touch navigation, future kiosks may allow customers simply to speak commands naturally while AI interprets requests and guides the transaction.

The document therefore predicts that accessibility regulations will eventually evolve to address this new interaction model.

Privacy Concerns

Perhaps the most important policy discussion in the document concerns privacy.

The authors compare home smart speakers with public kiosks.

Consumers knowingly install devices such as Alexa in their homes and accept cloud processing.

Public kiosks are fundamentally different.

Members of the public may unknowingly stand near an active speech recognition system.

The draft argues that this creates new obligations for transparency because nearby conversations could potentially be:

• captured
• transmitted
• analyzed
• stored

Consequently, people in the vicinity—not just the kiosk user—must be informed whenever speech recognition is active.

Universal Voice Recognition Symbol

One of the document’s most novel proposals is the creation of a universal industry symbol for speech-enabled kiosks.

The symbol would indicate that:

• voice recording is available
• voice recognition is available
• speech command is active

The proposal was reportedly developed with input from the Royal National Institute of Blind People (RNIB) to ensure that the symbol is:

• tactile
• visually obvious
• comfortable to touch
• easy to recognize

The document recommends using a single standardized symbol across all industries so users only need to learn one voice-interface indicator.

Illuminated Voice Status

Whenever speech recognition becomes active, the proposed symbol should illuminate using bright white lighting.

The illuminated indicator should:

• be visible before interaction begins
• be easy to locate by touch
• clearly indicate when the kiosk is actively listening

This serves both accessibility and privacy objectives.

Microphone Requirements

The Code proposes detailed engineering requirements for speech-enabled kiosks.

These include:

• permanently installed microphones or microphone arrays
• optional headset microphones
• automatic microphone muting
• activation only when a user enters the interaction zone
• automatic shutdown after the user leaves
• audible notification when microphones activate or deactivate

The interaction zone is defined as approximately 0.7 meters in front of the kiosk, with activation occurring after the user remains present for three seconds.

Environmental Requirements

Because many kiosks operate outdoors or in busy public spaces, the draft also recommends:

• beam-forming microphones
• advanced noise cancellation
• IP54 environmental protection
• 10-joule impact resistance
• resistance to sanitation procedures

These requirements acknowledge the harsh operating conditions of unattended public terminals.

Speaker and Wireless Audio Support

The Code extends accessibility beyond wired headphones.

Recommendations include:

• integrated amplified speakers
• automatic headset detection
• tactile volume buttons
• support for hearing aids
• support for wireless audio transponders
• continued compatibility with standard 3.5 mm audio devices

Wireless audio adapters are envisioned as a way for users to privately receive spoken guidance without physically connecting a cable.

Overall Assessment

This draft represents a significant evolution in kiosk accessibility guidance. While grounded in established standards such as the ADA, Section 508, and the Air Carrier Access Act, it looks ahead to a future in which AI-powered speech interfaces become a standard method of interacting with public technology. Rather than treating voice as a replacement for tactile controls, the document promotes a multi-modal accessibility model that combines touch, speech, audio, and visual interfaces so users can choose the interaction method that best meets their needs. It also places unusual emphasis on privacy and transparency, proposing visible and tactile indicators whenever voice recognition is active. If adopted broadly, many of these recommendations—particularly those involving speech recognition, microphone behavior, universal voice symbols, and AI-assisted interaction—could influence future accessibility regulations for public kiosks well beyond today’s ADA requirements.

Comments

One of the things mentioned when the COP was drafted was that consent (if needed) must be sought in a way that’s accessible. So for example if a microphone is used in a public space, there might be a state law (like BIPA in Illinois) which requires that consent be obtained before voice data is collected/used/stored. This sort of consent needs to be sought in a way that’s accessible, which is why the tactile microphone symbol was suggested along with a sensor, etc. Since the COP was drafted, even more states have enacted or are enacting privacy laws which cover biometric data, although it seems BIPA is still the strictest because it basically lets individuals sue even if no ‘harm’ occurred.

Also, depending on the kiosk’s use, there might be other laws with privacy regulations to consider, for example healthcare kiosks would also be covered under HIPPA (and I think there’s a law which covers financial services related kiosks like reverse ATMs).

For me, one of the main things I’d like the chance to explain to the US Access Board is that when it comes to self-service the ADA is not serving the people it was designed to protect. The ADA has too many holes that trained lawyers can exploit. We can’t even look to the lawsuits to see what might set a precedent because those also vary, depending on the circuit they’re filed in and what judge decides the case. For example some judges say having a member of staff on hand to help people with disabilities is acceptable, but others don’t see that as equal service so it’s not ADA compliant. They’ve got to update the ADA to include more specific guidance for self-service like the Europeans have with the EAA.

I still talk to people occasionally who think the ADA doesn’t apply to kiosks or self-service checkouts…

It’s 2026 — that was 2019

1. Voice Is No Longer Optional in Major Markets

2. AI Has Changed Both the Capability and the Threat Model

3. Privacy: From BIPA to a Patchwork of State Biometric Laws

• Colorado amended its Colorado Privacy Act (CPA) to include specific biometric data protections effective mid-2025. It requires informed consent before collecting biometric identifiers, a written retention and security policy, and prohibits the sale or trade of biometric data. Violations carry civil penalties up to $20,000 per violation.
• Comprehensive state privacy laws in multiple jurisdictions now explicitly include biometric data within their definition of “personal information,” subjecting kiosk operators to notice, consent, and deletion requirements that vary by state.
• Healthcare kiosks face overlapping obligations. HIPAA requires encrypted connections, biometric authentication, and role-based access control for patient data. As kiosks add voice capture for accessibility or telemedicine, operators must ensure voice data is treated as protected health information when linked to a patient record.

4. ADA Litigation Has Surged and Gone “Pro Se”

• 8,667 ADA Title III federal lawsuits were filed in 2025, with over 5,000 targeting digital properties—a 37% increase in digital cases.
• 40% of federal filings are now pro se (self-represented plaintiffs), many using AI to draft complaints and identify violations. This has expanded the plaintiff pool beyond traditional serial-filing law firms.
• Overlay/widget vendors have been penalized. The FTC reached a $1 million settlement with an accessibility overlay provider in 2025 for misleading compliance claims, reinforcing that bolt-on solutions do not satisfy legal obligations.
• The DOJ’s Title II rule (for state/local government) established WCAG 2.1 AA as the standard, with deadlines extended to April 2027–2028 depending on entity size. While this applies to government, courts consistently reference WCAG 2.1 AA as the de facto standard for private Title III litigation as well.

5. New Engineering and Operational Ramifications

6. The Global Standards Gap Is Closing—From Two Directions

• EU: The EAA is a product-safety-style directive with penalties up to €100,000 or 4% of annual revenue. It explicitly covers kiosks and mandates multi-sensory output.
• U.S.: No federal technical standard for private kiosks under ADA Title III, but litigation fills the gap. State biometric laws add a second layer of exposure.

Bottom Line

1. Accessibility is now mandatory in the EU and increasingly expected in U.S. healthcare and public accommodations.
2. AI makes voice interfaces workable but also makes voice data a biometric privacy liability.
3. State biometric laws have proliferated, so public kiosk voice capture requires explicit, accessible consent mechanisms—exactly what the draft CoP’s tactile illuminated symbol was designed to address.
4. ADA litigation is at record levels, fueled by AI-generated pro se filings, making structural multi-modal compliance (not overlays) the only viable defense.
5. HIPAA and Section 504 now impose specific digital accessibility deadlines on healthcare kiosks, adding regulatory teeth to the privacy and accessibility requirements your document outlined.