Description of the Technologies

Most user interfaces are designed to help users complete tasks. However, web security and privacy technologies intentionally introduce barriers to task completion. They require users to perceive more and to do more to complete tasks. Three examples of these technologies are passwords, CAPTCHA, and 2-Factor Authentication.

• Passwords are words or character strings used for authentication and/or for identity confirmation.
• CAPTCHA is a website widget, which prevents automated programs from submitting a web form intended for humans, by requiring humans to pass a test. Such tests:
  • present distorted text visually and/or aurally;
  • require users to enter that text into a field; and
  • require users to invoke a submit button.
• 2-factor authentication requires a two-stage process to verify the identity of a user. The user is required to have two of three of the following factors:
  • knowledge, e.g., password or PIN;
  • possession, e.g., mobile device or credit card;
  • inherence, e.g., fingerprint or voice print (via biometric device).

Challenges for People with Cognitive Disabilities

Web security and privacy technologies often block people with cognitive and/or physical disabilities who may not be able to:

• discern text they are required to enter and submit;
• recall text or instructions they have seen or heard;
• follow multi-step procedures.

The scope of the problem is vast because, for examples, people with disabilities:

• are prevented from purchasing goods and registering for services on the millions of websites that employ web security and privacy technologies;
• may circumvent web security and privacy technologies with insecure techniques/methods;
• may become so frustrated working through web security and privacy technologies that they relinquish their efforts, and thereby are thwarted from purchasing goods and registering for services;
• may be unable to become accustomed to a web security and privacy technology because there are multiple versions of it across websites;
• may be afraid to trust a web site, thus causing them to cancel a transaction.
  • Note: This is of particular concern for efforts to personalize web sites so they conform to users’ accessibility preferences. (E.g., users may be asked whether they trust a web site in order to pass such preferences.) See:
  • IndieUI
  • Global Public Inclusive Infrastructure (GPII)

Effect of memory impairments

Many people with cognitive disabilities:

• may have to look at or listen to text several times to copy or type it into a form field;
• may not recall steps needed to complete a procedure if an authentication session expires;
• may reuse passwords;
• may use simple-to-remember passwords, which are easy to guess/determine;
• may keep passwords insecurely, such as written on pieces of paper;
• may not recall where they keep passwords (which may be found by people who should not have them).

Some people with cognitive disabilities may not:

• be able to recall required text, such as a password or a PIN, or remember how to retrieve it.

Effect of impaired executive function

Many people with cognitive disabilities may not:

• complete a multi-step procedure for submitting text, such as a password;
• complete a timed procedure due to slowness in completing all steps;
• complete a procedure even if provided multiple opportunities to do so;
• enter characters in the correct order;
• enter characters correctly on the first try (resulting in being locked out).

Some people with cognitive disabilities may not be able to:

• retrieve required text, such as a password or a PIN;
• determine the purpose of a web security and privacy technology sufficiently or at all.

Effect of attention-related limitations

People with cognitive disabilities may not focus due to:

• frustration with time-limited procedures or presentations of digital security tokens;
• irrelevant instructions, such as CAPTCHA‘s “stop spam” and “read books”;
• presentation of multiple options, such as CAPTCHA‘s “Refresh”, “Listen”, and “Help”.

Effect of impaired language-related functions

Some people with cognitive disabilities:

• may have comprehension problems exacerbated by text or instructions presented in a non-native language.

Effect of impaired literacy-related functions

Some people with cognitive disabilities:

• may not comprehend the meaning of instructions related to web security and privacy technologies;
• may, when presented with words by CAPTCHA, be at a disadvantage due to lack of word recognition or comprehension.

Effect of perception-processing limitations

Many people with cognitive disabilities may not:

• read text at all because of the intentional distortion of it, a CAPTCHA technique;
• comprehend text that can’t be enlarged without additional distortion;
• understand text spoken in a computerized and distorted voice, a CAPTCHA technique;
• recognize characters if they do not form words, or are shown in different fonts/styles.

Some people with cognitive disabilities may not:

• understand the purpose of buttons such as CAPTCHA‘s “reset”, “listen”, and “help”;
• recognize functional elements, such as CAPTCHA‘s buttons, are clickable;

Effect of reduced knowledge

Some people with cognitive disabilities may not:

• recognize images, such as symbols or icons, of web security and privacy technologies;
• comprehend the meaning of rich media designed to be instructive.

Proposed Solutions

W3C Recommended Guidelines and Techniques

• Provide text alternatives that identify and describe the purpose of the non-text content.
  • See WCAG 2.0 Guideline 1.1.
• Turn off or adjust time limits, including allowing continuation of activity without reauthentication.
  • See WCAG 2.0, Guideline 2.2.
• Help users avoid and correct mistakes.
  • See WCAG 2.0, Guideline 3.3.
• Save submitted data for reuse after a user authenticates.
  • See WCAG 2.0, Technique G105.
• Encode user data as hidden or encrypted in a re-authorization page.
  • See WCAG 2.0, Technique G181.

Ease-of-Use Ideas

• Allow alternative authentication factors, such as:
  • location (e.g., user’s home or place of employment);
  • presence of a trusted family member or friend, who is detected, for examples, by a wearable biometric device or by a mobile device;
  • could be implemented consistently so that the interface is the same across web sites.
• Develop and use a consistent interface, such as common sets of vocabulary and iconography, across sites.
• Offer textual-password alternatives, such as swipe patterns or click-based graphical passwords.
• Provide security and privacy instructions and policies in plain language.
• Provide helpful feedback during web-form submission, such as explanations of:
  • why an entered password is insufficient; and/or
  • how to create a password that is easy to remember but hard to guess/determine.
• Set a high-security privacy option as the default, but ensure it is easy to use.
• Use a password-keeper app that is accessed biometrically, such as via fingerprint or voice print.

Alternative Web Security and Privacy Technologies

• Security tokens, some of which are hardware devices,can be used to make authentication easier. Security tokens are used instead of, or in addition to, other forms of authentication such as passwords. Security-token hardware devices:
  • include key fobs, rings, or small keypads;
  • can store and/or generate a digital signature, a PIN, or biometric data;
  • can transmit such data via a USB connector, RFID, Bluetooth wireless, or NFC.
• Keygen, an element of HTML5,can be used to simplify re-authentication. After a user has completed authentication using keygen, the user will be automatically authenticated for subsequent uses of a web site or service. Thus, there will be no need for a user to re-enter authentication information.
  • Keygen establishes a private-key and a public-key pair.
  • The keygen tag designates a key-pair field in an authentication form.
  • Upon form submission:
  • a private key is encrypted and saved locally; and
  • a public key is signed with the private key, and is sent to the server.
  • In subsequent authentication sessions, the server will either automatically retrieve the private key, or prompt the user to select it.
  • See W3C HTML5 Recommendation 4.10.12 The keygen element.
• Fast IDentity Online (FIDO), password-free standards for typical and two-factor authentication.
  • FIDO relies upon user authentication based upon a user’s device (e.g., phone, tablet, computer).
  • A user’s device registers the user, to a server, via a public key.
  • Upon a challenge from the server, the user’s device responds with a private key.
  • The device’s keys are unlocked by the user biometrically (e.g., fingerprint scanner) or by a button press, not by a password.
• Spam-free accessible forms, WebAIM, Utah State University, March, 2007.

CAPTCHA Alternatives

• CAPTCHA-less Security, Karl Groves, April, 2012.
• Inaccessibility of CAPTCHA: Alternatives to Visual Turing Tests on the Web, World Wide Web Consortium, November, 2005.
• Determine the time difference between when a web form is loaded and when it is submitted. If it is submitted quickly, which may be indicative of a spambot, discard the submission. Otherwise, keep it.
• A web-form honeypot that is:
  • an input field
  • hidden using CSS
  • labeled with a field name atypical of forms
  • clearly identified with instructions, for AT users, and for others whom have disabled CSS, not to fill it in
  • checked to determine if something was entered
  • used to reject a submission if something was entered

Notes

• I welcome your suggestions.
• I wrote this document as part of my work as a member of the W3C’s Cognitive and Learning Disabilities Accessibility Task Force. Other task force members contributed to the content.
• References in this document to “some people with cognitive disabilities” are to people with the lowest-functioning intellectual capacity, such as people with intellectual disabilities.
• This document:
  • is an expansion of CAPTCHA, Cognitive Disabilities, v1 (W3C Task Force).
  • offers answers to the challenges I described in Online Security & Privacy for People with Cognitive Disabilities, Part 1.
• The web-form honeypot, which I described in this document, will not work for popular websites because spammers will likely expend the effort to defeat it.
• For the latest official version, see the W3C Techniques Paper: “Web Security and Privacy Technologies“.

I am trying to evaluate the barriers of online security and privacy to people with cognitive disabilities. This work will help inform the effort of the W3C’s Cognitive and Learning Disabilities Accessibility Task Force to recommend standards on how to make online security and privacy more accessible.

Problem

I am struggling with how to go about this evaluation. It is a daunting task to come up with common barriers and solution recommendations across:

• the many end-user security and privacy techniques, e.g.:
  • passwords;
  • two-factor authentication;
  • biometrics; and
  • encryption
• the variety of platforms upon which the techniquesare implemented, e.g.:
  • operating systems;
  • devices;
  • web and mobile apps; and
  • messaging
• the different ways the many techniques within the various platformshave been implemented by the major players, e.g.:
  • Apple;
  • Google;
  • Microsoft; and
  • Open Source

Background

It is well known that people sacrifice security and privacy for the sake of convenience. Security and privacy techniques are too difficult to use, and thus are inconvenient. For people with cognitive disabilities, such “inconvenience” amounts to a significant barrier. (See my recent blog post about CAPTCHA for an illustration.)

It appears to me, from my research so far, that there is a lot of work on how to improve the security standards of information and communications technology (ICT) without much focus on the usability and accessibility of it. For example, I could not find even the terms “usability” and “accessibility” in the ICT Security Standards Roadmap of the International Telecommunication Union, an agency of the United Nations.

Improvement

Determining how to make online security usable for everyone must include people with cognitive disabilities. Doing so will mean that the related user experience will be designed to be as simple as possible. The more the experience is easy to use, the more everyone will protect their assets and privacy.

A piece of good news is that the Electronic Frontier Foundation is researching how to measure the usability of implementing secure messaging as part of its “Designing a Prize for Usable Cryptography”. I expect that work would be enough to help develop usability and accessibility-evaluation standards for online security and privacy in general; and to inform the creation of related recommendations for people with cognitive disabilities.

Solution?

I am working on a list of barriers, based upon functional limitations, which are common to end-user security techniques, and sublists unique to each technique. I am not a security and privacy expert. Thus the limitations I am considering are based solely upon my expertise in accessibility and cognitive disability, and what seems logical to me. (For an example, see my recent blog post about CAPTCHA.) I suppose that effort will have to suffice until a security expert, such as the Electronic Frontier Foundation, determines how to measure related usability.

Help Needed

I welcome comments with:

• suggestions about how to evaluate the barriers of online security and privacy to people with cognitive disabilities; and
• information about any effort to evaluate the usability and accessibility of online security and privacy techniques

Notes:

• See Stay Safe Online for online security-related instructions and information.
• No endorsement is intended or implied of the organizations and their efforts mentioned in this blog post.

As a member of the W3C‘s Cognitive and Learning Disabilities Accessibility Task Force, I agreed to review web-security technologies. I chose to begin with CAPTCHA. My first draft is below. The format I am using is the one I intend to use for future reviews. All the text is my own.

I welcome your feedback, additions, and/or revisions.

Definition

CAPTCHA is typically a website widget that prevents automated programs from submitting a web form intended for humans by requiring humans to pass a test. Such tests present distorted text visually and/or aurally; and require the form-submitter to enter that text into a field, and invoke a submit button. See http://www.captcha.net/

Problem

CAPTCHA often blocks people with physical and/or cognitive disabilities who cannot discern the text they are required to enter and submit. The scope of the problem is vast because, for example, people with disabilities are prevented from purchasing goods and registering for services on the (probably) millions of websites that use CAPTCHA.

People with Cognitive Disabilities May Not Be Able to:

• read CAPTCHA text at all because of the intentional distortion of it
• comprehend text that can’t be enlarged without additional distortion
• have the advantage of comprehending the meaning of words or images
• understand text spoken in a computerized and distorted voice
• complete the multi-step procedure for submitting the CAPTCHA text
• complete a timed CAPTCHA due to slowness in completing all steps
• understand the purpose of buttons such as reset, listen, and help
• recognize functional elements, such as buttons, are clickable
• focus due to irrelevant instructions such as “stop spam” and “read books”
• become accustomed to CAPTCHA because there are multiple versions of it

Alternatives

• Inaccessibility of CAPTCHA: Alternatives to Visual Turing Tests on the Web, World Wide Web Consortium, November, 2005.
• A Sliding Alternative to CAPTCHA?, L. Wroblewski, June, 2010.
• sweetCaptcha: Fun and Human Friendly Captcha

Notes

• Neil Milliken suggested I add that people may not be able to complete CAPTCHAs correctly due to sequencing problems causing them to input the characters in incorrect order. I will do so in my next draft.
• No endorsement of CAPTCHA or its alternatives is intended or implied.

The WC3‘s Cognitive and Learning Disabilities Accessibility Task Force received a presentation about a project called the “Global Public Inclusive Infrastructure” (GPII), from Gregg Vanderheiden, on March 31, 2014. Quoted below is a project description.

“The purpose of the Global Public Inclusive Infrastructure (GPII) is to ensure that everyone who faces accessibility barriers due to disability, literacy, digital literacy, or aging, regardless of economic resources, can access and use the Internet and all its information, communities, and services for education, employment, daily living, civic participation, health, and safety.

As our countries build out their broadband infrastructures to ensure that broadband reaches everyone, it is important that ‘everyone’ includes people with disability, literacy and aging related barriers to Internet use. We need to be sure that we don’t stop at just connecting people to the Internet – but that we also see to it that they can actually use it, and benefit from all that it has to offer.

The GPII would not create new access technologies or services, but would create the infrastructure for making their development, identification, delivery, and use easier, less expensive, and more effective.  Like building a road system does not provide transportation but greatly enhances the ability of car companies and others to do so — and provides an infrastructure that car companies themselves cannot do. The Internet is the infrastructure for general information and commerce. The GPII enhancements to the Internet would provide the infrastructure to enable the Internet to be truly inclusive for the first time.

GPII is a paradigm shift.  The GPII will, for the first time, introduce automatic personalization of user interfaces and user context adaptation based on user preferences.  Each information and communication technology (ICT) device will be able to instantly change to fit users as they encounter the device, rather than requiring users to figure out how to adapt, configure or install access features they need.  It also introduces a system of shared components and services to reduce cost, increase interoperability, and foster innovation.”

The “GPII is a project of Raising the Floor, a consortium of academic, industry, and non-governmental organizations and individuals.”

Retrieved from http://gpii.net (Published June 30, 2010).

Note: No endorsement of the Global Public Inclusive Infrastructure and Raising the Floor is intended or implied.

A new task force has been formed by the World Wide Web Consortium (W3C) to develop accessibility guidelines for people with cognitive disabilities. It is led by Lisa Seeman, a long-time expert and advocate. Task force members are well-known experts from all over the world.

I am a member, an “Invited Expert”. My current, primary responsibility is to create and manage volunteer research groups of people with disabilities and others. I participate in the weekly conference calls of the task force, which so far have consisted of brainstorming sessions, presentations, and organization by Lisa of the task force’s work. Our first teleconference occurred on January 20th of this year.

The task force is known as the “Cognitive and Learning Disabilities Accessibility Task Force (Cognitive A11Y TF)” of the Protocols and Formats Working Group, and the Web Content Accessibility Guidelines Working Group.

I plan to publish, to this blog, the information I can share about the task force’s work.