Three ways innovators can address gender bias in artificial intelligence

Leaders must be diligent in how they use the tech to ensure it helps, not hinders, the inclusion of people.

By Frank O'Dea Chief innovation officer, EY Ireland

EMERGING TECHNOLOGIES CONTINUE to overturn entire industries.

Artificial intelligence, or AI, is equipping the workforce with superpowers, helping to cure diseases and pushing the boundaries of space exploration.

In our company, it’s allowing people to complete tasks faster and more accurately than ever before, so they can focus on higher-value work for clients.

There’s tremendous opportunity in AI in Ireland, and today’s business leaders are seeing its benefits. However, as we push the boundaries of innovation, we need to do it responsibly.

Business leaders must be diligent in how they develop and use AI technology to ensure it helps, not hinders, the inclusion of people across gender, age, ethnicity and the whole spectrum of “difference.”

As chief innovation officers and other leaders consider issues facing business and society today, here are three ways organisations can take an inclusive approach to AI:

1. Address biases

In addition to solving complex problems quickly and accurately, AI is also being used to reduce human bias in decision-making processes. AI tools used in the hiring process, for example, can help organisations eliminate some biases to ensure hiring decisions are based on matching the desired capabilities of a potential employee with the innate skills of candidates.

However, what we need to remember is that if biases are embedded in the hiring tool, the AI will in fact reinforce this unintended bias in the hiring process. Therein lies the duty of humans in the process.

If we’re programming biases into AI technology, the future success and inclusive application of it cannot be fully realised. When developing AI, it’s critical to recognise biases, scrutinise algorithms and test the outcomes at every stage.

2. Diversify talent

Avoiding programming biases into AI technology and contributing to broader social inequalities, relies on the diversity of people within the organisation.

Why? There is a clear correlation between the lack of diversity in AI talent and distortions in some machine-learning outcomes.

Technology fields continue to be overwhelmingly male dominated; it’s estimated that only 18% of women hold top positions in AI disciplines globally, according to a study by Emerj.

Business leaders must ensure their talent pools are gender-balanced and representative of people and teams with the right mix of skills, experiences, education backgrounds and social, cultural and professional perspectives. This mix of diversity is critical at every stage of AI development.

3. Educate early

While it’s important for organisations to provide training programmes for employees to learn new skills and develop proficiency in AI, it’s also important to educate people at a grassroots level, by connecting with third-level institutions to ensure that this inclusive approach to AI is embedded in curricula from the beginning.

Focus must be directed towards educating future AI thinkers and leaders by creating pipelines for under-represented talent in AI and instilling the fundamental principles of using AI for the good of society.

Businesses should embrace the transformative potential of AI, and in doing so, take a responsible approach to it.

The business return is inherently obvious and as we will become smarter with our approach to AI, so too will the outputs of the technology.

After all, it’s how we use the AI, not the technology itself, that ultimately enables us to solve problems, not just for business, but for society and the greater good.

Frank O’Dea is chief innovation officer of EY Ireland.

Get our NEW Daily Briefing with the morning’s most important headlines for innovative Irish businesses.