Accelerating innovation with diversity in AI

Diversity is a driver of innovation 

Diversity and inclusion are increasing in importance with focus on its impact on many sectors, including business, academia, government and non-profits. Research has shown that increasing gender diversity has a positive impact on productivity, problem-solving, providing competitive advantage and innovation – all essential outcomes for tackling the great challenges of our time, economic growth, and tackle issues from health, food security, climate change and sustainable communities. 

This research highlights the potential value of team diversity as a practical tool for architecting decision-making processes,” said Harvard Business School Professor Francesca Gino. “That our decisions get sidetracked by biases is now well established. While it is hard to change how our brains are wired, it’s possible to change the context of decisions by architecting the composition of decision-making teams for more diverse perspectives.”

Source: whitepaper: Hacking diversity with Inclusive Decision Making

Another research conducted by McKinsey looked at 366 public companies across Canada, Latin America, the United Kingdom, and the United States. The findings indicate how diversity influences the financial performance of an organization. To summarize, “Companies in the top quartile for gender diversity are 15 percent more likely to have financial returns above their respective national industry medians”

Artificial Intelligence and the current challenges with data bias  

One of the tools that helps organizations accelerate their financial performance, obtain competitive advantage and boost innovation is Artificial Intelligence (AI). AI has proven examples of innovative products and services to create new markets as well as enhance existing markets. Amazon’s Echo is one such example. Other products that we use on a daily basis are Netflix movie recommendations, Facebook’s photo tagging, Uber’s price surge during rush hour and even the simple spam detector of our mailboxes. Business use cases of AI range from fraud detection to chatbots as well as complex medical applications such as breast cancer detection. 

However, increased use of machine learning, the most commonly used AI application, poses a challenge of amplifying bias in data, and brings forth the question of transparent and responsible systems. The ethical implications of AI systems have been a topic of recent discussions and framework that question any recommendation or judgement passed by the Artificial Intelligence application. 

Take the case of an Asian man’s passport being rejected by New Zealand’s passport checker stating that the “subject’s eyes are closed” 

This example confirms the fact that Artificial Intelligence applications need racial, gender, ethnically diverse teams to build inclusive products. Unlike traditional software development, AI needs arts and humanities to append the engineering teams. 

Artificial Intelligence and skills required

According to World Economic Forum’s Future of Jobs report, the top 10 critical skills for 2020 and beyond include complex problem solving, critical thinking creativity etc. 

These skills are well represented by women according to research. In addition, the research shows that women’s brains are more inter-connected between both left and right hemispheres and bring about a balance of the cognitive skills along with analytical abilities.

Human biases amplified by algorithms need the due diligence of diverse teams to analyze the data and the business processes, to monitor their impact and implement measures to reduce the impact of data bias in AI. As more jobs are automated, skills such as creativity, compassion, empathy, multi-tasking, patience etc. would increase in relevance and demand. Women are thus, better placed to bring these qualities to any AI project.

Another important point is that consumers are incredibly diverse when it comes to products transcending geographical boundaries as well as diversity increase in global population. For example, Deloitte’s Center for Consumer Insights reports “a shift over the last 50 years, from a homogenous US population to one that’s incredibly diverse and heterogeneous across all key demographics: race, age, generation, health, ethnicity, economics, and education.”

Homogeneous teams of AI engineers would overlook the needs of diverse population and different customer personas. Hence, a diverse team would be the first set of defense against algorithmic bias as well as weigh ethical aspects in the design, development and implementation of AI applications. Team members with different experiences and mindsets can avoid unintended biases, debate unfavorable outcomes to protected classes and design AI solutions that are transparent and responsible. 

Diversity and inclusion are imperative for any organization to be innovative and meet their current and prospective customers’ demands. Artificial Intelligence is central to innovative products and solutions that capture new markets and boost organizational productivity, reduce operating costs and introduce new revenue streams. However, AI comes with its challenges of bias in data, “black-box systems” and unexplainable algorithms. The need for a diverse team is crucial for designing and deploying ethical and responsible AI applications. Women bring a good mix of analytical skills added to cognitive abilities. They help drive innovation, question ethical implications of AI applications and help organizations build solutions that drive society towards a sustainable future.  

Read more on The Good AI

You may also like

Comments are closed.

More in BIAS