Unconscious (aka implicit) bias training used to be a rare request, but awareness of this important topic is increasing and now Unconscious Bias training is in the top 10 most requested trainings at Sunrise Well. Given we’re now navigating situations where people need more critical thinking than ever before, and especially as we’re seeing reports of aggressive behaviour moving offline and into the real world, smart workplaces are investing in moving beyond generic wellness into targeted topics with specified outcomes.
In this article, I explore the role of Unconscious Bias training in the modern workplace, what the science says about it, bias policies in Australia, and what to look for when engaging a trainer.
What is unconscious bias?
Like it or not, Unconscious Bias permeates the way we live and how we view the world. And by it’s very nature being unconscious, even the most self-rated fair-minded people have at least some unconscious bias interrupting their decisions and actions. Also known as implicit bias, unconscious bias is the automatic, unintentional stereotyping of people based on group characteristics like age, gender or race, which shapes decisions and judgments without the person being aware of it.
The big question on everyone’s lips in workshops I run on unconscious bias is, ‘how did it get here?’ The truth is, there’s no single defined answer. In reality, unconscious bias stems from a complex mix of mental shortcuts, pattern-making and familiarity, and social conditioning. So the question of whether it’s a nature or nurture problem is hard to answer: it’s both.
There are many types of bias in our lives, but some of the more common ones are:
affinity bias
confirmation bias
halo effect
age and gender bias
anchoring
The way unconscious bias shows up in the workplace begins before someone even starts at a job. Unconscious bias has been found in recruitment, onboarding, performance reviews, promotions, and leadership. It also shows up in sports, educational opportunities, health care, and basically any aspect of society that involves people.
Join The Sunrise Well Newsletter
What the science says
Implicit cognition research grew from social psychology in the 80s and 90s. Research from the time found that people form and use mental associations automatically, and often they’re not aware of this automatic activity. The Implicit Association Test (IAT) (Greenwald, et al 1998) measured how quickly people pair concepts. For example, they looked at how quickly they paired a social concept, like family, and a positive or negative word. The theory was that faster pairing signals stronger associations.
Harvard has been running Project Implicit for many years now and millions of people have taken the IAT online, feeding more information into the data to improve understanding of unconscious bias.
One of the most notable drawbacks of this work is that there are differences in people’s scores from one sitting to the next. General understanding of the IAT now is that it can be a useful tool for reflection, but shouldn’t be used in any kind of diagnosis or labelling of someone’s bias.
One thing that’s interesting to note though, is that unconscious bias shows up for everyone. This is not a matter of intelligence. You can’t rationalise or pontificate your way out of bias. It is implicit in nature, and it affects people across all areas of society. What it does show though, is that awareness and reflection is important.
The benefit of Unconscious Bias training is not to immediately effect change in decision-making. The aim is to reflect, gain new perspectives, practice new ways of approaching a situation, and implement structural change with this greater awareness. The training acts as an important piece in the process, to kick-off the space of reflection and testing, however it is usually what happens after the training that is the most impactful.
AI inherits our assumptions
Large language models derive their information from vast amounts of human-written text, and, like it or not, that text carries the stereotypes, gaps and historical imbalances of the societies that produced it. For example, it’s well-known that models link the word “doctor” with men and “nurse” with women, because that pattern has been reflected in the data. The AI model simply reproduces it.
Humans place a lot of trust in machines, sometimes to our own detriment. At this point of the AI journey, people are over-trusting the output by AI models without enrolling their critical thinking skills to check and verify whether the information is correct. Beyond that though, there is also an increased risk of confirmation bias in AI. Confirmation bias means accepting AI answers that match what we already believe. This is affected by the way we word our questions, and the
While this bas in AI is now well-known, it doesn’t mean it has been fixed. The best way to combat AI bias is to train and ensure that the people using it have critical thinking skills, and the autonomy to trust the information given to them.
In the news: when bias is automated
Mobley v. Workday: In 2023, a job applicant alleged Workday’s screening tools discriminated by race, age and disability. Workday argued that age discrimination law excludes job applicants, but this was rejected in court in 2026. it’s important to note that at this point, we’re talking about allegations, not confirmed activities, but it’s still worth considering the role automated tools play in reinforcing decisions which are biased.
In terms of repercussions, many AI-related cases are still in front of the courts, however the U.S.-based Equal Employment Opportunity Commission (EEOC) has stated that saying “the algorithm did it” is not a valid defence. Organisations need to be on the front foot about how AI is used in it’s recruitment processes and other people-management processes.
What effective Unconscious Bias training looks like
When looking to engage an outside provider to deliver unconscious bias training, it’s worth having a few questions up your sleeve to make sure they’re the right fit. You can ask the trainer to share how the workshop demonstrates and provides reflection on:
critical-thinking and decision-making skills, not just awareness.
practise with real scenarios
methods for questioning AI outputs
opportunity to create processes for hiring and promotion
opportunities for leadership involvement.
When looking at how to question an AI output, it’s important to build in a pause before moving forward. Here’s a quick framework you can take with you and share with your teams today:
PAUSE
Pause: Take a moment to pause and check in with yourself. What’s your first reaction? Does the answer match what you already believe? (possible confirmation and automation bias at play)
Ask for the basis: Where is the answer from? Check the sources. Verify the key data independently.
Uncover who’s missing: Whose experiences, voices, or information might be missing or under-represented here?
Stress-test: Prompt the AI tool to argue the opposite. Rephrase the question and rerun to check if the output is completely different. Compare the output to a secondary source.
Ethical ownership: Ultimately, the outcomes fall under human ownership. Ensure accountability by owning the results. Record your reasoning, especially in hiring, performance, and promotion so that the decision can be traced and understood.
Where to start
A lot of people don’t know where to start when it comes to reducing unconscious bias in their workplace, so if you’re feeling a little overwhelmed by it, you’re not alone. I suggest reviewing your policies with an understanding of bias in mind. Pay particular attention to your recruitment and promotion process, but also remember that bias can show up across the entire work spectrum in many different forms. Unconscious bias training should form part of your plan to reduce bias in the workplace, but it should be part of a comprehensive program. If you’re feeling overwhelmed, you’re welcome to reach out for an obligation free chat. I’d love to share more about the unconscious bias training programs available at Sunrise Well.
Ruth Kent is the Founder of Sunrise Well, and has delivered unconscious bias training to organisations across Australia. With qualifications in Psychology, Public Health, and Biomedical Meditation Therapy, and training as a professional health coach, she brings her 27 years of facilitation experience to the forefront with engaging and interactive learning sessions. Ruth is a professional member of Health Coaches Australia and New Zealand and Meditation Australia, and is also an active member of the Positivity Institute’s Collective+ for workplace coaches and consultants. Learn more about Ruth here.
FAQ
-
Unconscious bias training helps people recognise the automatic assumptions and mental shortcuts that can influence their judgments about others. The most effective programs go beyond awareness to build practical skills, such as structuring decisions, questioning assumptions and tracking outcomes.
-
It depends on the design. Research shows awareness-only training has little lasting effect on behaviour, and a large 2019 meta-analysis found that changes in implicit bias did not reliably change biased behaviour. Training that teaches practical strategies, includes follow-up and is paired with changes to processes is more likely to yield positive outcomes.
-
No law specifically requires it. However, employers have a positive duty under the Sex Discrimination Act to take reasonable and proportionate measures to eliminate sex-based discrimination and harassment, and law firms advise that bias training can be part of meeting that duty. What counts as reasonable depends on the organisation’s size, resources and risks.
-
There’s no research-backed ideal frequency. One study found that the effects of awareness training faded within two weeks, which suggests one-off sessions aren’t enough. A common approach is an initial workshop followed by regular refreshers alongside bias-aware processes and policies.
-
Unconscious bias is a mental pattern, while discrimination is an action or decision that treats someone unfavourably because of a protected attribute such as age, sex, race or disability. Bias can lead to discrimination, and under Australian law intent isn’t required, so a decision driven by unconscious bias can still be unlawful.