The Truth about Artificial Intelligence in the Hiring Process
I believe that companies and organizations should not integrate technology or artificial intelligence (AI) into their hiring process in order to prevent discrimination. Although technology and AI do have benefits, I believe the shortcomings do not outweigh these benefits and are simply too risky for a business to take. Throughout the New York Times article, Dr. Ifeoma Ajunwa highlighted many negative connotations associated with AI and how it would have a detrimental effect on employment relations and the hiring process.
One of the largest noted shortcoming of implementing AI technology into the hiring process is the lack of transparency that exists with the use of AI.[i] We currently do not know how AI is making its decisions as we, as outsiders, have no way to truly know what the machine has been instructed to do.[i] This doesn’t only leave room for organizations to discriminate against individuals, but it permits them to do it without public knowledge. Currently an organization can claim that certain criteria are part of cultural fit, even though these criteria may be highly discriminatory towards particular individuals.[i] In addition to this, it was found that only certain individuals were receiving ads on sites such as Facebook to ensure that people similar to the an organizations current employees would be ‘recruited’ by the organization.[i]
AI is usually being implemented to mitigate the issue of human error with regards to discrimination; however, as noted by the articles author, in a published paper, AI receives its data from humans making the process somewhat redundant and just as discriminatory as humans.[iii] This counter-productivity has occurred on several occasions when it was proven that AI was in fact discriminating against specific individuals that had something in common.[iv] As a result of this in 2017/2018 the Ontario Human Right Tribunal received close to 5000 accusation applications, 40% of which were found to be discriminatory.[v]
I do not think that AI is advanced to the degree that will allow it to be integrated into the hiring process. AI has not been developed to point that it can be interchangeably used with humans.[ii] As a result of this, the technology or AI that is used is not capable of detecting company-candidate fit to the same degree that humans are able to detect. The use of technology and AI could also be rendered useless if the data is from humans as it could be discriminatory towards a particular group of individuals.[iii] It is critical that the relationship between the employer and potential candidates remains civil, respectful and professional. To the best of its abilities, an organization should not be discriminating against its potential candidates through AI technology as it could result in a case being filed against their organization. That is not to say that in years to come this technology will become more advanced permitting it to be an aspect of the hiring process.
[i]Ajunwa, I. (2019, October 8). Beware of Automated Hiring. Retrieved February 2020, from https://www.nytimes.com/2019/10/08/opinion/ai-hiring-discrimination.html
[ii]Balinggan, G. (2019, January 1). Difference Between. Retrieved February 2020, from http://www.differencebetween.net/science/difference-between-artificial-intelligence-and-human-intelligence/
[iii]Ajunwa, I. (2018, December 3) The Paradox of Automation as Anti-Bias Intervention. Retreived February 2020, from DOI:10.1093/jlb/lsz003.
[iv]Holmes, A. (2019, October 8). AI could be the key to ending discrimination in hiring, but experts warn it can be just as biased as humans. Retrieved February 2020, from https://www.businessinsider.com/ai-hiring-tools-biased-as-humans-experts-warn-2019-10
[v]Tribunals Ontario. (2018). Social Justice Tribunals annual report 2017/18. Retrieved February 2020, from http://www.sjto.gov.on.ca/documents/sjto/2017-18%20Annual%20Report.html
Drew Barden 
This article piqued my interest because of a conference I attended last year, Economics in the Digital Age. Speakers business and technological backgrounds all came to the same conclusion: automation in the workplace is inevitable. Although I strongly agree with your points on the dangers of AI and automation on employment relations, I think that we cannot avoid it so we must focus our efforts into regulating the bias that AI contains. I agree with your point about discrimination in the AI process. I see job postings on my Facebook and LinkedIn profiled based on posts I’ve favoured, but I hadn’t considered the background checks that may be involved. This closed loop creates exclusion by targeting specific company goals. (https://www.nytimes.com/2019/10/08/opinion/ai-hiring-discrimination.html ) There must be systems in place to understand these exclusions as discriminations. What would this system check for to prevent discrimination?
I also agree with your point about transparency. I am most concerned about my privacy and not knowing what is being done with my information. Forbes released an article on AI job hiring, where cognitive and emotional features of candidates were assessed using their facial movements and vocal tone, then compared to employees who have succeeded in the roles. (https://www.forbes.com/sites/bernardmarr/2019/05/29/artificial-intelligence-in-the-workplace-how-ai-is-transforming-your-employee-experience/#a94357353cec ) It is scary to think this personal information could be floating in a digital cloud. However, this analysis could be used to recommend candidates to other roles, expanding opportunities for employment that may not have previously existed. Do you think the consequences outweigh the benefits?
How do you feel about the article’s suggestion of unionization to protect the interests of employees? I think employees should take an active role in ensuring fair hiring platforms for companies to solidify their needs in the workplace. Ajunwa explains how unionization could determine criteria for job fits and protect applicant data. (https://www.nytimes.com/2019/10/08/opinion/ai-hiring-discrimination.html) I think this would be an effective way to ensure the fairness of the system. Having employees advocate for their hiring would create a fairer process.
Finally, I can’t talk about automation without thinking of job turnover as robots and AI replace jobs in low-skill fields. Especially in manual labour, companies have shifted to automation to increase productivity, increasing the customer experience. (https://www.ft.com/content/c4bf787a-d4a0-11e9-a0bd-ab8ec6435630)I ) I think in order for this to positively affect employment relations, discarded employees must be retained to value-added activities. Automation should not create unemployment in the economy or else the productivity gain will cancel out.
Overall, I agree that we must understand how automation creates discrimination so that we can implement systems and unions to correct it. The increasing role of technology in the workplace can improve employee relations if it contributes to human-added capital, instead of discouraging it.
LikeLiked by 1 person
Hi there!
First of all, I would like to thank you for shedding light on various aspect of my blog post, it is much appreciated. I think whether or not the consequences outweigh the benefits really depends on the organization. I personally would not like myself to be assessed for a position based on facial movements and vocal tone. I’d prefer to be chosen for a position because of who I am and the skills I possess. If an organization is using this information as a large part of their decision criteria in their hiring process than I would say that the benefits do not outweigh the consequences. However, if it would simply be considered as a small aprt of the decision criteria and they see this characteristic as beneficial to the position then perhaps the benefits do outweigh the consequences.
To address the other part of your comment I question the unionization to protect the interests of employees, I am again on the fence. The topic of transparency plays a large role in this area as often employees and candidates are very much in the dark during the hiring process and don’t always know what exactly is going on and by what criteria are they being assessed by. If this unionization would create a more transparent process both internally and externally to potential candidates. I would see this as very beneficial and would make everyone less skeptical about a hiring process that they may be in.
Yes of course, turnover, employee retention and employability all become huge aspects to this ordeal. I definitely agree that employees must be retained to value-added activities in order to not create more harm than good. However, I’m sure many companies would struggle with this and individuals who are being replaced by AI or other technology would find themselves without a job and potentially a change in career if they are no longer needed in a field that they previously specialized in.
If you have anything to add, I would love to hear about your further insights on this topic.
LikeLiked by 1 person
*Test*
LikeLike
Drew,
I agree with you that discrimination continues to happen with the involvement of AI in the hiring process despite the efforts made to reduce the amount of discrimination happening. Recently, in my workplace, the company started to use a software called Checkster to complete background checks on potential employees. The immediate feedback on the software was that, as a company, we lose the interaction with the person who is providing the reference. The change to this practice has removed, in my opinion, a very valuable piece of the reference checking process. I guess it could be argued that the personal touch for which people may value, may also be the culprit of the discrimination – albeit not on purpose. We have bias’ that are out of our control based on background, experiences, etc.
I agree with the comment above from 17maw10 that automation in the hiring process is inevitable. It is better to have some oversight on how it plays out, rather than placing it on the back burner and dealing with it only when it is too late.
Very insightful Blog. Thank you!
Tiffany
LikeLiked by 1 person
Hi Tiffany!
I definitely agree and I like that you both have shed light on the alternate side of the topic. It is definitely a difficult topic to discuss and I truly do feel that there are valid points to argue both sides. Without the way our society it progresses AI will inevitably be further integrated into both the hiring process and many other aspects of our day-to-day lives. I am not super familiar with Checkster; however, from what I understand I agree that it really does lose that personal touch. As stated in my blogpost I do see the value in having a real human performing these as we are able to analyze in a different way than AI. As you said, this may be partially biased due who we are as people; however, it is a better analysis. If you personally were involved in a hiring process would you prefer that it had AI integrated into the process or a process without AI? Or maybe a combination? I feel as if a combination with a more prominent presence of humans could be beneficial. Thank you for insightful comment, it is much appreciated!
LikeLiked by 1 person
This article gained my interest as it is an opinion that is not highlighted often in today’s AI world. How can a company understand it’s employees if they are chosen through an AI process? I agree that not only is this an ethical issue that must be resolved, but it could end up hurting companies productivity in many factors including work culture, peer support and guidance with upper-level employees.
I think clarifying areas in business that are strengthened with AI is key to understanding why certain aspects of business to not conform to the convenience and ease of AI. I would recommend including a few examples within that aspect.
Although AI is most likely conformed to less biases than the average human, I believe that human interaction within the hiring process is something that can never be eliminated. Great article.
Matt
LikeLiked by 1 person
Hi Drew,
I used the same topic to share my opinions on the matter. After reading your opinion piece, its interesting to see the similarities in viewpoint – positively speaking. I think that when it comes to AI or automation in the workplace the ultimate goal is obvious. Which is; save time money and in turn be more efficient at whatever job duty needs to be carried out. Do I think this works in some cases – yes. Do I believe that it should be implemented? In some cases yes and some no. The variables are not the same across the board for every workplace situation. Do I think that we should be undercutting jobs to humans? No, but again there’s variables that come into play here.
I will say that having worked in a high volume staffing firm back in my day, it was pretty clear that the employees in the office showed up late, left early, took long breaks, socialized a lot, and worked the system (if it was unionized) lets say by exaggerating specific situations in order to get an extended leave. Did they need that leave with pay? No, but they did it because the politics and process of it would take forever to get resolved, and that was common knowledge.
Also, the number of potential candidates that had to be interviewed per day was staggering. It was impossible to get through 100 candidates, take notes, and come up with some sort of assumption we know what kind of worker they are at the end by one interview that took 5 minutes. Therefore, I think that it doesn’t really matter if its AI or humans operating behind the interview process. Humans are more than capable of discriminating intentionally and non-intentionally, and the principle (in some cases) would still be the same for both humans and AI which is; how do we hire the best by judging someone based off their resume. The by product is still to review and assess written professional credentials on paper by asking this one question “what type of worker is this person”? We would still have to assess the verbal interview and the resume; as humans. The same goes for AI just without the face-to face portion. They are essentially cutting out one phase of the process as really it doesn’t matter what the person looks like, its their consistent work ethic taken into account. I cant attest to that because I was the interviewer. I did not judge someone by how they dressed, the color of their skin or what ethnic background they came from. It was all from whats shown to me on paper and their job profile. So what’s the main concern then? I think that by creating a computerized system, we as humans are assuming it would be a flawless process or the software is supposed to run without errors. If we as humans are creating these issues; by using AI its supposed to fix it right? Again, that’s just the assumption we are under considering how technology has ultimately created a more effective solution to many issues and in many different environments in our society.
At the end of my opinion piece, I agree with Dr. Ajunwa. Which is to keep it simple and regulate the use of the technology by implementing strategic employment guidelines it must follow (just as humans would have to do), analyze the functionality of it and come up with effective solutions on how to maintain a stable platform that could potentially be very beneficial in the long run.
Looking forward to your response if interested.
Cheers,
Amanda
LikeLiked by 1 person