Most people don't want a company AI to do the hiring, firing, or promoting

[midian182]

A hot potato: Not only could AI take your job, but it might also decide whether you get another one. More companies are using the technology for hiring and other workplace decisions, but almost two-thirds of Americans would not want to apply for a position with an employer who uses AI to make hiring choices. And even more are opposed to AI making the final hiring decision.

The recent meteoric rise of generative AIs such as ChatGPT looks set to revolutionize the technology landscape. But for all the good it can bring, automation always comes at the cost of jobs. Research by bank Goldman Sachs claims these systems could replace a quarter of work tasks in the US and Europe, the equivalent of 300 million jobs.

Companies aren't just using AI to replace human workers - as the creatives at one company found out – it's also being used extensively for recruitment and other HR-related purposes.

But, much like talking to someone who's using generative AI, most people don't want an artificial intelligence being part of the hiring process. A recent Pew Research study that surveyed 11,004 Americans found that 66% would not want to apply for a job with an employer who engaged in this practice. Just 32% said they would apply, while the rest were unsure. Even more participants, 71%, were opposed to AI making the final hiring decision.

As is usually the case when companies start incorporating AI in their workflow, most claim using it in the hiring process speeds everything up, such as screening and contacting candidates, giving human HR workers – those who haven't been replaced – more time for other tasks.

Elsewhere in the survey, 62% of people believe AI will have a major impact on workers over the next 20 years, but just 28% think it will affect them personally. Moreover, only 13% believe AI will help more than hurt workers, while 32% think it will equally hurt and help, and another 32% say it will hurt more than help.

Looking at how AI can be used for monitoring & evaluations and facial recognition, the only two areas people were in favor of it being used were monitoring workers' driving behavior as they make trips for the company (43% in favor) and automatically tracking the attendance of employees (45% in favor).

The results show most participants are against employers using AI, especially when it's used to decide who gets hired, fired, or promoted. One of the fears people have is that AI can be biased, which we've seen before.

"AI as it is usually applied today looks for specific words or qualifications that often miss the whole picture," one respondent told Pew. AI can also allow "structural biases based on race or socioeconomic status to persist unchallenged," he added.

Others pointed out that AI would not take into account factors that don't appear on a job application.

Conversely, the minority who favored AI said it wouldn't have the biases found in some humans, such as ageism, and it could look past non-traditional work histories.

Permalink to story.

https://www.techspot.com/news/98450-most-people-dont-want-company-ai-do-hiring.html

 

[m4a4]

On one hand, it could be programmed to be unbiased and only look at skill and merit (which also means not doing any image/facial recognition beyond checking against online databases so you don't hire a wanted criminal).
Taking into account age and other similar factors in context isn't something I'd consider biased (as long as it isn't an extreme).

But on the other hand, we all know that it will be biased and have some sort of quota to fill or list to check off. The only companies that will start out using this will be the bigger ones. And we all know that they love to pretend they aren't biased.
And that will be on top of AI not being able to tell all that well if there is deception (something a competent human would have an easier time doing) among other "organic" things (like social situations).

If anything, they could be used to help filter candidates in a factual way so that a human can have a closer look.

 

[James Ryan]

It's amazing how easily people will bend over and take it.
Ai could be amazing for humanity.
The problem is that it's currently being fueled by capitalism for capitalism.
Nothing else but PROFIT...

 

[RudyBob]

That's tough. If AI could weed out the folk that desperately want a job and will agree to everything the employer asks then almost immediately forget that they went to that company and asked for a job and signed a contract verbal or otherwise and start saying things like this company sucks and I deserve ______________ then I am all for AI hiring

 

[RudyBob]

It's amazing how easily people will bend over and take it.
Ai could be amazing for humanity.
The problem is that it's currently being fueled by capitalism for capitalism.
Nothing else but PROFIT...

Hooray for Capitalism!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! If Techspot wasn't capitalist you could not post

 

[brucek]

Ultimate accountability has to rest with a human.

An AI that flags signs of poor decision making, especially illegal ones, could be a nice supplement to the process and very valuable to those responsible for supervising it (I.e., not the immediate hiring manager, but the VP of HR who is ultimately accountable for how the last 1,000 people were hired.)

For a long time to come either should be able to override any AI concerns or suggestions but merely having the AI report as part of the discoverable legal record along with the supervisor's note as to why concerns were not considered valid or overridden could be a nice balance for all parties against favoritism, discrimination, and I suspect by far the most common of all, just plain stupidism and lazyism.

 

[BuckarooBonzaii]

If I remembered, some were fearful of computers in it's early stage decades ago and now the whole world uses them. AI is now where computers were once was, as AI will not go away as it is a part of our evolution.

 

[Tanto]

Facial recognition! But is AI smart enough to differentiate between bad attitude and resting ***** face?

 

[LetTheWookieWin]

From what I’ve seen, there has been a continual progression towards using metrics (of often dubious value) to try to quantify performance and productivity and remove human judgment. I’m not going to make a blanket statement that metrics are inherently bad, but from my experience there is a real need for managers who are willing to provide direct and honest performance feedback (and in the correct venue). Using AI and metrics won’t fix bad management.

 

[waclark]

From what I’ve seen, there has been a continual progression towards using metrics (of often dubious value) to try to quantify performance and productivity and remove human judgment. I’m not going to make a blanket statement that metrics are inherently bad, but from my experience there is a real need for managers who are willing to provide direct and honest performance feedback (and in the correct venue). Using AI and metrics won’t fix bad management.

I think metrics are absolutely important in evaluating a person's performance. The challenge is that often the metrics being used to evaluate you are not always explained to you. I've always done well in jobs where the metrics were pretty obvious, such as you have to sell a certain dollar amount of equipment. Or you need to product a certain number of widgets per day. It's when you get into the "works well with others", "goes above and beyond" etc that it is harder to be consistent and objective about the evaluation. Managers will overlook things with people they like and will magnify small issues with people they don't like. This is where AI might be of use and that is building an evaluation based on observable and measurable statistics, not some feel-good bullshit.

 

[LetTheWookieWin]

I think metrics are absolutely important in evaluating a person's performance. The challenge is that often the metrics being used to evaluate you are not always explained to you. I've always done well in jobs where the metrics were pretty obvious, such as you have to sell a certain dollar amount of equipment. Or you need to product a certain number of widgets per day. It's when you get into the "works well with others", "goes above and beyond" etc that it is harder to be consistent and objective about the evaluation. Managers will overlook things with people they like and will magnify small issues with people they don't like. This is where AI might be of use and that is building an evaluation based on observable and measurable statistics, not some feel-good bullshit.

Just to be clear, I'm not saying metrics are unimportant. The issue I'm trying to articulate is that metrics are often viewed as this panacea that will result in better performing teams, without addressing the need for effective leadership. Many metrics are poorly designed, unrealistic, and are easily "gamed". For example, if the metric is "build x components per period of time" or "reduce the cost of x by 10% annually", you could easily meet your metric at the expense of product quality (and not be penalized). Similarly, if the metric is "complete x work requests per period", I have seen staff simply log every little thing they do as a work request to inflate their numbers to the point where the overhead of logging these tickets actually reduced throughput.

My point is that I personally don't see the lack of metrics or AI processing of said metrics as the final solution when a lack of competent management is the real root cause for most business failures.

While I get your point about managers who have personal biases involved, I'd still rather work with a human than with an algorithm that judges me 100% on metrics alone, without the ability to see anything beyond the numbers (or judge whether the numbers made sense in the first place).

Anyway, that's my perspective. Thanks for the reply.

 

[monkeylove]

One problem with AI replacing people is that businesses use the tech in place of employees to be more productive and thus make more goods and services to sell, but the ones expected to buy those are the same employees.