With the right support, could AI help us narrow equity gaps?

As I head to Mobile World Congress in Barcelona, there will be lots of talk about AI (Artificial Intelligence) and the role of technology. Almost two-thirds of CEOs surveyed in a recent EY CEO Outlook Pulse see AI as a “force for good” suggesting enormous potential for AI to level the playing field, including identifying bias and creating new, more equitable opportunities for people to thrive and grow.

As AI is rapidly transforming global business, the workforce and raising the bar of what's possible, it is more important than ever to look at AI's potential through a Diversity, Equity & Inclusiveness (DE&I) lens and the capability it has to help us narrow inequity gaps. Of course, there are very real risks of algorithmic bias in AI systems that can lead to unjust outcomes, disproportionately affecting minority ethnic groups, but if implemented properly, AI could help us address current DE&I challenges for the better.

How the use of AI tools can close gaps to inequitable access

Evidence is emerging that indicates AI workplace tools are helping lower-skilled workers up the productivity ladder, so they begin to catch-up with higher-skilled workers. One study by the National Bureau of Economic Research showed that access to AI increases productivity by 14% on average, including a 34% improvement for novice and low-skilled workers.

By supporting workers with new tools and levelling the playing field in this way, traditional workforce barriers can be broken down, opportunity opened up, and the talent pool democratized, encouraging a more inclusive workforce for many underrepresented groups whether that be in terms of education, neurodiverse needs, socio economic background or geography.

AI can show us patterns that may be in our blind spot, which can negatively impact certain groups or demographics. Greater visibility into these patterns can ultimately support more equitable performance evaluations, promotions, resource allocation and project assignments.

AI can also analyze data related to employee engagement, satisfaction and feedback to help leaders tailor initiatives that provide a sense of belonging and inclusiveness. This data can help customize DE&I actions based on specific needs and experiences.

With the seismic shift brought about by AI, numerous new professions and job opportunities are on the horizon. One significant outcome of this shift will be the emergence of new job categories and roles that will necessitate a blend of creativity, critical thinking, numeracy, and attention to detail, all without the requirement for complex in-person interactions with others.

At EY, we’re working to equip professionals with the knowledge and skills to leverage AI technologies effectively and responsibly to enable seamless AI-people collaboration to achieve extraordinary outcomes. One example is a TextHelp product we’ve piloted at EY called Read&Write. The technology supports employees who think, work and learn differently by offering employees a choice in how they understand and are understood. 

For example, users can hear content read aloud rather than read, which supports information processing. They can use advanced writing features to remove the worry over the mechanics of spelling, use organizational tools to summarise information easier, and support memory challenge. The tool can also collect digital content from multiple sources into one document, with automatic source referencing.

This has proven particularly useful for neurodivergent employees, as they often think, learn, process and behave in ways that are more unique to the typical population. With this support comes the benefits of out-of-the-box thinking, creative solutions and greater productivity.

There’s also an important AI application in recruiting. EY’s Talent Attraction & Acquisition team is exploring how AI algorithms can help us interact with more candidates from a wider array of backgrounds, without having to manually sort through thousands of resumes and CVs.

Our AI-driven Candidate Assistant helps job seekers find the open roles at EY that are best suited to their background and preferences, but it won’t preclude them from applying to any other job that they may be interested in.

Another example we’ve employed is the Success Factors Job Analyzer. This technology helps us ensure that our job descriptions – tens of thousands of them – are reviewed for potential gender bias and revised accordingly. It is embedded into our core technology and drives change in our hiring mindset and practices.

As we adopt more technology examples like these, there’s endless potential to make our workplaces more equitable and inclusive. That’s only if we use our human qualities to show the technology right from wrong. That means constant checks and balances, and education and awareness of the risks for everyone using the technology. Here are three ways to implement those checks and balances:

1.    Ensuring all employees are building inclusive behaviors as well as AI skills. Better education is the best way to tackle AI bias and to make the most of the technology. At EY, we’re offering e-learning courses to all of our 400,000 people, including Inclusive Leadership for All and Upstanding Interventions (guidance on when to intervene in situations of inequity). We also offer practical tips for teams to ensure everyone contributes to group discussions.

2.    Setting up equity guardrails. While policymakers work to develop a shared understanding of issues and risks to close the AI trust gap, organizations have their own work to do. EY formalized its AI principles, which include accountability, security and privacy, transparency and explain-ability, fairness and inclusivity, and professional responsibility. These principles form a key component in the rollout of AI learning to all EY people, so everyone has equal access to AI opportunities.  

3.   Considering diversity in AI design, testing and procurement – To uphold more ethical and inclusive outputs, organizations can influence how AI is designed, tested and adopted. Examples include requiring representative data sets and input from individuals from diverse backgrounds and abilities, along with embedding equity checks and balances into AI algorithms to produce more inclusive results.

AI holds the promise of augmenting our potential to deliver better results that benefit everyone. But we know people working with AI is much more powerful than AI alone. That’s because people can contextualize intelligence, cultural nuance and lived experience. Implementing AI technology with human, DE&I mindset is how we can uncover more diverse perspectives, address inequities, drive greater innovation, productivity and, ultimately, revenue. It will also help to foster more personalized, equitable careers and an inclusive culture where everyone belongs.