How AI Will Impact DEI Efforts
Like diversity, equity & inclusion (DEI), AI has its champions and skeptics, yet both are gaining significant traction in accounting firms. DEI initiatives aim at ensuring equity and opportunity, while AI is revolutionizing processes. In essence, both represent transformative forces, shaping the progression of the profession. Here’s how AI will impact DEI efforts based on what we’ve seen so far, and how can we use AI to streamline our work without falling prey to its downsides.
How AI Will Impact DEI: The Good
If used properly, AI can enhance and improve a firm’s efforts to be more equitable and inclusive. One of the ways this is happening is through the use of resume scanning tools to reduce unconscious bias in candidate selection and hiring. By removing information that indicates race, gender, age, and other identifying factors, hiring managers are more easily able to analyze each candidate’s skills and experience without bias. This is important when multiple studies have shown that employers seeing identical resumes were 50% likelier to advance an applicant with a stereotypically white name versus those with more Black-sounding names.
Language scanning tools like Textio can be used to analyze content and recommend changes to make wording more gender neutral, so it appeals to a more diverse audience. In addition to obvious gender changes like using “businessperson” rather than “businessman,” there are more subtle changes that can make a big difference. Certain words and phrases appeal more to men (like “competitive,” “dominant” or “leader”) while others will attract more women (“support,” “understand” and “interpersonal”).
AI can also scan materials for screen reader compatibility and color schemes for those with visual impairments, while speech recognition tools can convert spoken words into text for neurodivergent learners. If your goal is to reach as many people as possible without giving preference to any one group, AI can really help.
How AI Will Impact DEI: The Bad
Of course, there is another side to this coin. Users should always remember that AI learns from existing data, which can reflect societal and individual biases and inadvertently perpetuate stereotypes. On top of this, 75% of AI programmers are white men, which can lead AI to favor that group and unintentionally punish others. How does this play out in real life? Autonomous cars find it more difficult to “see” people with darker skin tones and are more likely to run into them. Also, programs trained on American English have been known to penalize people with regional accents, non-native English speakers and those with speaking impairments.
Another factor to remember is that AI will generate false information, referred to as hallucinations. Not to be confused with simply pulling the wrong data, hallucinations are when AI makes up information from whole cloth. I recently wrote an article on an attorney who used AI to prepare for court and discovered that the case law it pulled didn’t actually exist. In other cases, it has claimed the world record for crossing the English Channel on foot is held by Christof Wandratsch who did it in 14 hours and 51 minutes on August 14, 2020, and the Golden Gate Bridge was transported across Egypt in October of 2016 for the second time. Not only is it lying, but it is also pretty specific and creative about it.
Embracing the Good and Avoiding the Bad
It is easy to over-rely on a tool that makes our lives easier, but it is not always the best route to take. And ignorance is not an excuse. So how can we get the most from AI without becoming a victim of its bias or hallucinations?
- Determine where data comes from: Regardless of the tools you use, research where data is sourced, what features are included to avoid bias, and how often it is updated. Most tools have a process for reporting issues too, so don’t be shy about using it when you find it is feeding you inaccurate information.
- Review results for built-in bias: Understand how bias can sneak in and impact results so you can routinely audit your AI queries for prejudiced language or questionable outcomes.
- Ask if this is a job for AI or a human: Over-relying on AI can lead firms to replace necessary human input and interactions with machine learning. Educate yourself, and your team, on when AI tools are appropriate and when they are not.
When we get down to it, how AI will impact DEI will depend on how we use the technology. Like any tool, it can make life easier, but must be used with caution and precision.
About Bonnie Buol Ruszczyk
With 25 years of marketing experience, Bonnie Buol Ruszczyk helps professional services firms develop unique strategic marketing that will help them reach their growth and awareness goals. This can take many forms including serving as a fractional or part-time CMO, creating strategic marketing plans for the firm as a whole and the niches it serves, helping hire and manage internal marketing teams and much more.
Welcome to CPA Growth Trends — your source for information, insights, tools and best practices to drive growth within an accounting firm.
with Danielle Reynolds, Business Development, Manager with Whitley Penn
A business developer’s day involves a myriad of activities from external meetings with business owners and referral partners to scoping calls for initial client connections.