Computers are supposed to help humankind by helping ease the mental load, but they can also work against us with bias found in employment software, home-loan applications and even pizza delivery.
Subscribe now for unlimited access.
$0/
(min cost $0)
or signup to continue reading
Shahriar Akter, Associate Professor at the University of Wollongong, has been delving into the dark side of artificial intelligence (AI) with his research previously assisting in the $112 million class-action against the Federal Government's over their Centrelink Robodebt bungle.
The latest research he's worked on (published in the academic Journal of Business Research) uncovers algorithmic bias in mortgage applications by banks, recruitment biases by various human resource organisations and even search engine bias through Google.
"AI is going to affect every facet of our lives in this digital/algorithmic world," Professor Akter said.
"I do believe that algorithms enable us to make evidence-based decisions and they can transform our lives in the future ... [but] when data driven AI-algorithms are used in public administration and are inextricably linked to people's livelihoods, especially in the context of welfare payments, there has to be a clear management capability and real accountability."
Illawarra resident Umme Hani claimed she experienced bias from recruitment agencies when she moved to Mount Druitt in Sydney's west, with not a single call-back or acknowledgement from dozens of applications sent to recruiters over a three-year period.
Mrs Hani said within the first two months of moving to Wollongong she received several job interviews, having sent out the same resume.
"I was mainly applying for business administration jobs," the UOW PhD student said. "I read Shahriar's research article and it ... makes me think there must be a reason I never got call backs."
A northern Illawarra career and HR expert (who wished not to be named) said HR companies had been using algorithmic software for 20 years which he believed beneficial though "only a small part of the recruitment process". However, he did admit there was bias.
"I know employers exclude some people based on many factors as they are entitled to; but it is easy for people to complain or claim discrimination when they are just not suitable," he told the Mercury.
"Many Sydney based companies exclude people from outer Sydney, suburban and regional areas as the travel into the city causes attendance and punctuality issues.
"Many Sydney North Shore and Eastern Suburbs companies do not employ 'westies'."
I know employers exclude some people based on many factors ... many Sydney North Shore and Eastern Suburbs companies do not employ 'westies'.
- Illawarra Career Expert
The expert said "covert discrimination" happened every day but was hard to prove, and said the best way to beat the system was to ensure your resume included key words found in the job description criteria and ensure you will be able to fulfil the role's duties. (scroll down for his tips)
"A café role that required a person from 6am to 2pm had one applicant who had school-aged kids and expected to be able to bring the kids to work and then leave for 45 minutes to take them to school each day [during her shift]," he said.
"[Software] predominantly works by searching certain key words and phrases that match the programmed phrases.
"All industries have their own language, terms, acronyms etc and these are used as filters to identify who is suitable (many people have a simple non specific resume, with no cover letter addressing the criteria and apply for multiple roles and they are not successful)."
The research co-written by Professor Akter also uncovered flaws in an array of areas like tax assessments, insurance premiums, credit card subscriptions, surveillance, airport passenger screening, hotel booking and take-away delivery.
"Mortgage algorithms generally identify [single mothers, and Aboriginal and Torres Straight Islander people] as less profitable customers in making mortgage decisions," Professor Akter said.
"A bank evaluates mortgage applications and determines the creditworthiness of applicants based on an algorithm using its historical mortgage approval data.
"The historical applications that were approved by previous employees indicated bias against certain social groups such as young people, blue-collar workers, and single-female applicants."
He said bias is generated when algorithms reflect they might reflect historical or social biases due to poor data quality or inefficient algorithmic design - resulting in unjust, unfair or prejudicial treatment of people.
If organisations (especially banks) become make their algorithms more "transparent", Professor Akter said that would help people understand the reasons they didn't get a job or declined a loan while also helping make the process more fair.
"An algorithm enables service providers to arrange predictive maintenance of their high-value equipment, improve performance, reduce cost and downtime; customers can enjoy a more personalized, consistent and engaging service experience," he said.
"In the [future] algorithms will not replace humans rather they will augment human decision making in various service innovations."
Have you experienced bias applying for a job, or a home-loan, or trying using the web for takeaway delivery? We'd love to hear from you - EMAIL desiree.savage@illawarramercury.com.au
CAREER EXPERT'S TIPS
- Have a specific resume for the role you are applying for (generic resumes are too general and the algorithms will exclude people who are not specific)
- Address the advertised criteria via a specific cover letter (a specific cover letter allows you to tell the employer how you meet the advertised criteria and are the best candidate)
- Ensure that industry, workplace employment experience terminology and lingo is specific to the role that you are applying for (algorithms pick up key words and support your application)
- Make yourself stand out. Do not just rely on the online applications but call the Advertiser / Recruiter / HR Manager where the job vacancy is and speak to a person so that you can present yourself as a person and not a number (this will show that you are keen, motivated and being resourceful)
- Follow up all applications with a phone call or email to check the progress (again this is the same as 4 and also gives you an opportunity for feedback if you are not successful).
- Be proactive as advertised jobs only make up 30% of the job market. The hidden job market is the still the best way to get a job (canvassing, going in person, calling up, sending in expression of interest emails, going into a shop when you see a sign on the door stating vacancies or in trade based offering to come in and do a short stint of work experience).
The Illawarra Mercury newsroom is funded by our readers. You can subscribe to support our journalism here.