Numbers and Nuance: Being Aware of Bias

The 2022 Diversity and Inclusion Working Group survey revealed a clear message: Institute members are seeking to understand unconscious bias to further enhance the integrity of their work.

In this article, we explore unconscious bias, its impact on the actuarial profession, and strategies to mitigate it.

As the “Becoming an Actuary” page on Actuaries Institute’s website says, “Actuaries see and shape the future”[1]. This foresight stems from more than our technical capabilities; it is anchored in our ability to question assumptions and challenge the status quo. However, the subtle yet powerful force of unconscious bias can impede this critical ability.

Professionals, including actuaries, make three billion decisions each year.[2] We often use the term “actuarial judgement” to refer to the decisions we make, drawing on the knowledge gathered and synthesised from our experiences. Unfortunately, human cognitive processes are resource-intensive, so a little thing called “heuristics” is employed.

Heuristics are mental shortcuts to improve the speed and efficiency with which humans make decisions. Whilst heuristics can speed up our problem and decision-making processes, they can also introduce errors and biased judgements. As actuaries we have been taught the skills and ethical frameworks to ensure that the assumptions, reliance and limitations of our technical work are clear and mitigated against bias, but there are unconscious biases that can creep into many other facets of our careers.

What is unconscious bias?

Unconscious bias refers to the automatic, often negative, stereotypes and attitudes we form about individuals or groups without conscious awareness. These biases are deeply ingrained and often shaped by our upbringing, cultural background and personal experiences. While we may strive for impartiality, these unconscious beliefs can subtly influence our perceptions, decisions and behaviours.

In the actuarial profession, the consequences of unconscious bias can be far-reaching. From data selection and model development to risk assessment and decision-making, these biases can distort our understanding of reality, leading to inaccurate predictions and inequitable outcomes.

In data science and our development of artificial intelligence (AI) models, there are many examples of where AI has proven to perpetuate underlying biases. One example is the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) AI algorithm used in US courts to predict recidivism risk, which was incorrectly biased towards people of colour (POC).[3]

Another example is Amazon’s AI recruitment agent. Developed in 2014, it was found to be biased against women as the ten years of data it was trained on reflected the high-performing Amazon employees at the time, i.e., men.[4]

Beyond the numbers and models, there are many examples and studies of biases present in the workplace that actuaries should be aware of:

  • Affinity bias: Hiring or promoting people who share similar backgrounds, interests or experiences. For instance, a manager might favour candidates from the same university or with similar hobbies.
  • “Halo Effect”: Overvaluing a candidate based on a single positive trait, such as a qualification, overlooking other capabilities (or lack of capability).
  • Name bias: Subconsciously associating certain names with particular ethnicities or socioeconomic backgrounds, influencing perceptions of competence.
  • Gender bias: Rating female employees lower than male counterparts for similar performance, often attributing male success to ability and female success to luck.
  • Age bias: Assuming older workers are less adaptable or technologically proficient, leading to unfair performance evaluations.
  • Stereotyping: Making generalisations about individuals based on their membership in a particular group (e.g., assuming that women are better at event organisation than men, or that men are better than women at maths).
  • Microaggressions: Subtle and often unintentional behaviours or comments that exclude, offend or demean people from marginalised groups.
  • Parental status bias: Assuming employees with children are less committed or available for work.
  • Weight bias: Making assumptions about an individual’s work ethic, intelligence, or health based on their weight. [5]

 

From a diversity and inclusion perspective, recognising unconscious bias is crucial. Unconscious biases can create barriers for individuals from diverse backgrounds and limit representation, leading to inequitable outcomes, the stifling of innovation, and damage to an organisation or profession’s reputation. By acknowledging and addressing unconscious bias, actuaries can create a more inclusive environment where people feel valued and respected, ultimately leading to increased diversity of thought, improved decision-making, and stronger business outcomes.[6]

How can we mitigate against our own unconscious bias?

The very first step to mitigate your own unconscious bias is to be aware of it. It may be shocking to know that many people (myself included) hold prejudices we don’t truly believe. This is because unconscious biases are formed over time through exposure to societal stereotypes, media portrayals, and personal experiences.

If you are curious about biases that you hold, you can head to Project Implicit, an organisation that researches “implicit social cognition” and whose mission “is to educate the public about bias”. You can take one of their tests, particularly the ones on Social Attitudes to see if you hold any unconscious biases.

Awareness on its own will not create change. While acknowledging unconscious bias is a crucial first step, true progress requires ongoing action. Here’s how actuaries can take an active role in mitigating bias in their professional lives:

  • Embrace a growth mindset: Recognise that our brains are adaptable, and we can learn to overcome bias.
  • Seek diverse perspectives: Actively solicit input from colleagues with different backgrounds and experiences.
  • Challenge assumptions: Question the “why” behind decisions and be open to alternative explanations.
  • Practise self-reflection: Regularly assess your own decision-making processes for potential bias.
  • Advocate for inclusion: Support initiatives that promote diversity and inclusion within your organisation and the profession.

 

By adopting these strategies and fostering a culture of continuous learning, the actuarial profession can evolve into a more inclusive and equitable space and spread that evolution to the industries and organisations that we work within.

This, in turn, will enable actuaries to leverage diverse perspectives and enhance the integrity of our work, promote equity, and build a more resilient and sustainable future.

What steps will you take today to challenge your own unconscious biases?

If you’d like to explore a specific aspect of unconscious bias in more detail, or the role of diversity and inclusion in the profession, please contact Bianka at bianka.grange@quantium.com.au or another member of the Diversity & Inclusion Working Group.

References

[1] Actuaries Institute. (2024). Becoming an actuary. Actuaries Institute Website. https://www.actuaries.asn.au/becoming-an-actuary

[2] Larson, E. (2017). Don’t fail at decision making like 98% of managers do. Forbes. https://www.forbes.com/sites/eriklarson/2017/05/18/research-reveals-7-steps-to-better-faster-decision-making-for-your-business-team/?sh=5e8342d440ad

[3] Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias. ProPublica. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing

[4] Dastin, J. (2018). Amazon scraps secret AI recruiting tool that showed bias against women. Reuters. https://www.reuters.com/article/us-amazon-com-jobs-automation-insight/amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK08G

[5] Randall, R. (n.d.). Workplace bias. Eddy. https://eddy.com/hr-encyclopedia/workplace-bias/

[6] Carucci, R. (2024). One more time why diversity leads to better team performance. Forbes. https://www.forbes.com/sites/roncarucci/2024/01/24/one-more-time-why-diversity-leads-to-better-team-performance/

CPD: Actuaries Institute Members can claim two CPD points for every hour of reading articles on Actuaries Digital.