What 18 months of being homeless taught me about bias
In August 1999 I was living at my parents house but our relationship was bad. Towards the end of the month they told me I had to leave and I had three days to move out. Heading into my second year of college and without much in the way of reserve funds I took to sleeping on friend’s couches, sleeping in student lounges at the college and working solo overnight cleaning jobs where I could finish my work quickly and catch some shuteye.
For eighteen months I relied on the kindness of friends (and some strangers) for meals, places to sleep and help paying tuition.
Years later, while working with a mission agency who served the homeless, I came to realize one of the reasons I was able to exit out of my situation was of my network. The people I knew and had grown up with were all in positions to offer me a place to stay, extra food in their fridge and were friends of people who could arrange to have my tuition deferred.
Not everyone has this fortune and in Pragya Agarwal’s, Sway: Unraveling Unconscious Bias, she offers a more fulsome perspective: being male, white and educated put me in a better position to propel forward than someone who didn’t have those traits.
In Canada, nowhere is this more apparent than in the federal penitentiary intake system. As part of the process of a new prisoner being admitted to federal prison, they are first housed at an assessment centre where a decision is made on which institution they will eventually be transferred to and the same assessment will be used when considering parole[2]. The assessment is a points based system which sees individuals who are married, have had stable employment, no history of mental health issues scored favourably[3], however, these are all scores which have been imperially proven to favour white inmates and reduce the chances a black or indigenous inmate will be sent to a better prison and ultimately have access to less rehabilitation programs[4].
Bias at this extreme is easy to spot, the Offender Intake Assessment is published on the internet[5] and data about Canada’s prison system is publicly available for organizations like the John Howard Society to comb through and comment on. But as Agarwal points out, bias often exists in the shadows where it isn’t as easily weeded out. During the height of the COVID-19 pandemic, Zoom became the de facto tool for video conferencing. The problem for people with darker skin, however, was when they would enable Zoom’s virtual background feature it would remove their head[6].
But this isn’t even a new problem, Agarwal writes. The default settings for digital cameras favour light skinned people because the settings are programmed to emulate film – which in turn was built to expose white people well[7]. She describes a process where skilled photographers will ensure they take special care to fine tune their camera in order to properly expose darker skinned subjects.
But really, shouldn’t technology help us weed out bias? After all, technology is not a human and even AI needs to be programmed to learn, so teaching technology to not be biased should be pretty simple.
Apparently not.
Amazon developed a tool to scan résumés and promote suitable candidates to hiring managers while discarding applicants who didn’t fit the bill. The software was trained on over 50,000 keywords that researchers had developed based on past successful hires. It was a machine learning tool meaning as it worked, it would get better at finding the best applicants[8]. But something sinister began to happen. Slowly the algorithm began to screen out candidates who had attended all-women colleges, the word women or “women’s” like in “captain of the women’s netball team.”[9]
After being reported by Reuters, Amazon confessed they had to abandon the software because the bias programming had been implanted “too deep” into the system.
This deep programming has a long history. Sheryl Sandberg, who at the time was the COO of Facebook, recalls a meeting she had at a blue chip firm in New York City. She was the only female in the meeting and at one point she asked her male colleagues where the bathroom was. None of them were able to tell her because they had no idea. One wasn’t even sure if a woman had ever been in the meeting room they were using[10].
Agarwal says the solution for combatting bias in technology lies with the tech companies themselves. Firstly they need to be transparent in their design and programming and exclude any sort of micro targeting which would leave out women, non-white people and other groups that have historically been the victims of bias.
[1] Pragya Agarwal, Sway: Unraveling Unconscious Bias (London: Bloomsbury Sigma, 2020)
[2] Siobhan O’Connell and Ayobami Laniyonu, “Race, Gender, and Risk Assessments in Canadian Federal Prison,” SAGE Journals, accessed March 14, 2024, https://journals.sagepub.com/doi/full/10.1177/21533687231153993
[3] John Howard Society of Alberta, Offender Risk Assessment (2000), accessed March 14, 2024, https://johnhoward.ab.ca/wp-content/uploads/docs/OffenderRiskAssessment_2000.pdf
[4] Tom Cardoso, “Bias behind bars: A Globe investigation finds a prison system stacked against Black and Indigenous inmates,” The Globe and Mail, accessed March 14, 2024, https://www.theglobeandmail.com/canada/article-investigation-racial-bias-in-canadian-prison-risk-assessments/
[5] “Offender Intake Assessment,” accessed March 14, 2024, https://publications.gc.ca/collections/collection_2021/scc-csc/PS84-123-2009-eng.pdf
[6] Megan Rose Dickey, “Twitter and Zoom’s algorithmic bias issues,” TechCrunch, accessed March 14, 2024, https://techcrunch.com/2020/09/21/twitter-and-zoom-algorithmic-bias-issues/
[7] Ibid., 386
[8] Ibid., 366
[9] Ibid., 366
[10] Sheryl Sandberg, “Why we have too few women leaders,” TED, accessed March 14, 2024, https://www.ted.com/talks/sheryl_sandberg_why_we_have_too_few_women_leaders
About the Author
Mathieu Yuill
While raindrops on roses and whiskers on kittens may be a few of Julia Andrews' favourite things, here are a few of mine: Talking to strangers, Learning about what you do for fun, Conversation over coffee. I own a marketing and communications company in Toronto, Canada called Leading With Nice. There are a lot of names I could have given the company but a trusted friend encouraged me to name it that because I really value the humanness in us all. Bah - this is starting to sound like a horrible LinkedIn post. So whatever, let's have coffee. I'd love to hear about what you do for fun!4 responses to “What 18 months of being homeless taught me about bias”
Leave a Reply
You must be logged in to post a comment.
Thanks Matthieu. As I read your post, I heard you process your own ‘advantage’ in your unhoused journey as compared to others–an important thing to be aware of (Thanks, Sway). But also as I read your post, I thought about the bible verse that says, “Whenever we have trouble, God comforts us. Because of that, we ourselves can comfort other people. When they have any kind of trouble, we can comfort them…” and I wondered if your later volunteer work with a homeless agency was the result of your own unhoused experience–perhaps including a different perspective on ‘homeless people’ as a group and some of your previously held bias that shifted through your experience? That’s perhaps more than you might want to blog about…so we can talk about it in Washington as well!
Your story is a great one to underscore the power of privilege. It’s so important to note this, that “some” have the resources to climb back up just based on who they are and that they have a safety net. We wouldn’t be where we are today if we didn’t have help in a down payment for a house, or buying a minivan when pregnant with a 3rd child in ministry. It’s hard to imagine ever hitting rock bottom, and yet many other’s are never given those chances. I wonder if the best way to bring light to implicit bias is to be honest of our own? Thanks for sharing Mathieu.
Thanks for sharing a part of your story. It reminded me of the old saying “try walking a mile in another man’s shoes” but you put another layer in there that you realized, some people don’t even have shoes to put on. This book has helped reinforce that none of us are off the hook when it comes to implicit bias. Thanks for your post!