Sign up to our newsletter and be the first to receive new episodes and the latest news in DE&I straight to your inbox.

  • News

DiverCity Podcast in the news

City AM blog: AI, diversity, and the pursuit of broadened thinking

AI and the need for a diverse approach are increasingly central to corporations’ longevity

The intersection of science and diversity is at a crucial stage. As artificial intelligence and machine learning become increasingly cemented within the financial services industry, there comes a deeper need to understand how biases can become ingrained within the AI-related tools being used within organisations. What needs to be done to overcome these challenges and can the power of data and AI help create diverse and high-performing teams?

Rana Gujral, CEO of Los Angeles-based Artificial Intelligence company Behavioral Signals, and Anna Tan, a Singapore-based consult who specialises in corporate transformation through the lens of neuroscience and social psychology, came together on Episode 5 of Series 11 of DiverCity Podcast to discuss their experiences of working on the cutting edge of technology. 

AI is a global industry with predicted revenues of £236 billion in 2021. From Silicon Valley, Israel, and East Asia, hotspots for AI research and design have dethroned traditional seats of research power, making advanced computing a worldwide industry. As such, listening to voices from across the world is vital to develop an understanding of what is happening within the space and how it will impact efforts to promote diversity both in the UK and further afield. 

We are all loaded with biases – both unintentional and otherwise – and this creates a problem when designing AI tools and software. If the same biases are coded into the decision-making process of an AI application, then they are disseminated further, perpetuating and reinforcing those biases. So for AI researchers and companies, coming to terms with their own biases is vital to ensuring that they are not replicated in their products.

If it is not bias-free, then the AI tool is “useless,” says Gujral, adding that a bias-free learning model is essential to the longevity of the product, and subsequently to the business that operates it. 

“There’s a silver lining in this whole equation that, at least in the AI world, if you’re going to build a biased AI model, essentially you will fail as a business in the long term,” adds Gujral.

But this is not to say that creating a bias-free system is the ideal solution. The idea of ‘good and bad’ biases is integral to our own decision-making. This applies when we discuss diversity because making conscious efforts to observe our decision-making process can result in better AI products.

Diverse thinking, both independently and as a corporate team, allows you to consider a broader range of factors, and the same applies when developing AI tools. 

“The more diverse you are,” says Gujral, “the more perspectives you have, the more worldviews you have – your decision-making is more sophisticated and more advanced and better.”

This is a view reflected by Tan, who being based in Singapore, experiences the crossroad of corporate research and diversity in everyday life.

As a specialist in the science of diversity, Tan is uniquely placed on the leading-edge of corporate trends. As data science and AI increasingly permeate the industry, her work delves into the notion of biases and prejudice and how corporations can use neuroscience to foster a more inclusive mindset.

“Dr David Rock did some fantastic research,” explains Tan, with a model called SCARF – a theory that describes how we respond to certain situations.

“SCARF stands for Status, Certainty, Autonomy, Relatedness, and Fairness – and this is all hard-wired in every human being. When you’re triggered negatively in any of these five domains, you go into the threat state and then you’re either fighting for your life or you’re running away from fear.”

This understanding can be applied to a corporate context, especially in the way leaders often repeat the same perspectives without acknowledging a broader range of views, often resulting in confirmation bias.

Tan says it is vital that leaders break that pattern by becoming more conscious of why they think and work as they do.

Tan’s work is truly international, encompassing various markets in the Asia-Pacific region, which results in her encountering various cultural attitudes, sometimes stemming from biases around background and education. 

In Asia, for example, Tan says that there is “almost a snobbery of academic achievement.” This happens, she says despite there being no correlation between high achievement in a corporate role and high achievement academically.

Meanwhile, Western expats in Asia are “expected to have more of a worldly view” and their education is “deemed higher than some of the locally attained qualifications,” says Tan. This is despite the fact that it is more academically challenging to get into a local university in Singapore.

These views regarding educational achievement are clear examples of biases that have become ingrained within corporate attitudes.

It is clear from the experts that the proliferation of AI and the pursuit of a more diverse way of thinking are not mutually exclusive. There are opportunities for leaders to challenge their own thinking and to develop technology that integrates broad and varied ways of dealing with corporate challenges.

You can listen to the full podcast here