Standing Firm on Equality
At UserWay, we stand firm on equality in all forms. In an extension of our quest to build an internet that is accessible for all, we are beginning to identify and implement ways that our solutions can promote equality beyond the disability space.
The same AI-powered technologies we created to identify and remediate accessibility violations in a website’s code can be used to identify potentially offensive language or even blatantly discriminatory text. In the same way, our solutions can automatically identify WCAG violations, it can now flag text that may be racist, biased or socially unacceptable.
Our intention is not to censor or silence, but to make content owners aware of problems that exist on their site so they can resolve them. This is a natural next step for our technologies, and we hope it will assist web teams in identifying and removing racist or socially unacceptable content swiftly while also shedding light on problems they may not previously have recognized.
Introducing our AI-Powered Content Moderator
UserWay’s AI-Powered Content Moderator helps keep your website free from discriminatory, biased and racially charged language. We do this by intelligently flagging content for review. Similar to emoji skin tones modifiers, the Content Moderator offers nuanced alternatives to words and phrases that may be considered sensitive. Nothing will be deleted or removed without your approval.
These capabilities have recently been rolled out to all of UserWay’s customers as an integral part of UserWay’s AI-Powered Accessibility Widget.
Improving Your Company’s Content Policy
UserWay’s AI-Powered Content Moderator opens the door to establishing digital information policies with a strong core of diversity, equality and inclusivity. Spark awareness within your organization and inspire real changes in your communication policies by critically evaluating the content you allow on your site.
Tapping into our powerful Content Moderator is an important part of your organization’s journey towards authentically inclusive content that engages people regardless of their race, ethnicity, age, gender, religion or sexual orientation.
How it Works
UserWays AI-Powered Content Moderator scans your website’s text, both static and dynamic, and flags potentially objectionable words and phrases and any language that may promote stereotypes or prejudice. This includes texts that could be considered racist, sexist, anti-semitic, homophobic, xenophobic, violent, intolerant or otherwise offensive.
You can then review the Content Moderator’s suggestions and choose the ones you’d like to accept. Additionally, you can edit the suggestions to seamlessly flow with your site’s content.
Safety, Not Censorship
The Content Moderator cannot be used as a censor. It does not automatically reject or remove content. It simply locates and highlights cases of non-inclusive and biased language for review and replacement by site administrators.
Civil rights advocates, led by the Anti-Defamation League (ADL), have increased pressure to ensure websites are carefully moderated and keep hate speech off of their platforms.
Recent calls for repeal of Section 230 of the U.S. Communications Decency Actmay expose online publishers to future legal action for defamation based on opinions or reviews created by platform users. UserWay now helps you mitigate the legal risk of both ADA- and ADL-related violations.
Bring Your Brand up to the Standards of Your Audience
Consumers are now demanding that brands take a proactive approach to inclusiveness, and companies are responding.
The makers of Aunt Jemima, Redskins candy, Uncle Ben’s, Eskimo Pie, and other products have announced plans to rebrand their products to rid them of racial and ethnic stereotypes. At the same time, marketers and technology companiesare working to make their content, copy and computer code racially and gender-neutral.
Many have gone so far as to eliminate phrases that connote that “black” is bad (blacklist, black sheep, etc.) and “white” is good (whitelist, white hat, white lie, etc.), and other terms that hinder inclusiveness.
UserWay’s AI-Powered Content Moderator will flag these, along with thousands of other words and phrases, for your team to review.
Handling Legacy Content
Content, particularly when it exists on large platforms, is often created by interdisciplinary teams working in tandem. It is a challenge to reflect back to the team members involved about any potentially sensitive, prejudicial, or divisive language.
A content moderation project attempting to screen legacy content on a 1,000-page website (a standard size for large businesses) can take up to six months and two full-time employees to conduct a review and make modifications, and significantly more for media sites with half a million pages or more. Yet, they may only catch a fraction of the potentially biased language on the site.
UserWay offers a fast and less costly alternative. Simply add the UserWay AI-Powered Accessibility Widget to your site and gain the benefit of the AI-Powered Content Moderator.
The Current Landscape
Content Moderation, like accessibility remediation, is done in the face of constantly evolving and rapidly-changing standards. Content Moderation, in particular, responds to changing cultural understanding and awareness.
Anonymized statistics shared by sites already using the Content Moderator allow the tool’s underlying dictionary to constantly evolve. UserWay’s team evaluates and curates the terms gathered from this data to ensure each one’s accuracy and relevance.
Early samples of data from over 500,000 websites show that the majority of issues detected concern gender and racial equality, as noted in the table to the right.
Early data samples of UserWay's AI-Powered Content Moderator
Bias Type | % of Sites |
---|---|
Sites with racial bias | 52.45% |
Sites with gender bias | 24.32% |
Sites with age bias | 11.75% |
Sites with racial slur | 4.77% |
Sites with disability bias | 3.06% |
Sites with religious bias | 1.29% |
Sites with ethnic bias | 1.08% |
Sites with sexual orientation slur | 0.92% |
Sites with sexual orientation bias | 0.35% |
Sites with disability slur | 0.08% |
Sites with disease bias | 0.01% |
The Power to Decide
Site administrators can now decide whether to accept a suggestion, replace the word or phrase with one they choose themselves or allow the original language to remain unchanged. Site administrators have full control, allowing the decisions to stay internal and ensuring that all changes make sense and preserve the original meaning of the message.
As a worldwide community, we should never tolerate systemic racism and socially irresponsible conduct. Like many individuals and businesses, UserWay not only takes a stand against this type of speech and behavior, but provides solutions to help institute change in an actionable and proactive way.