Key part is that AI is suspected of down-ranking folks by age (ADEA = Age Discrimination in Employment Act)
> The Court has provisionally certified an ADEA collective, which includes: “All individuals aged 40 and over who, from September 24, 2020, through the present, applied for job opportunities using Workday, Inc.’s job application platform and were denied employment recommendations.” In this context, being “denied” an “employment recommendation” means that (i) the individual’s application was scored, sorted, ranked, or screened by Workday’s AI; (ii) the result of the AI scoring, sorting, ranking, or screening was not a recommendation to hire; and (iii) that result was communicated to the prospective employer, or the result was an automatic rejection by Workday.
Age discrimination is a huge issue and I've experienced it firsthand. Places want to hire younger people because they're more apt to work longer hours for less pay. It's going to get worse as people who got into the web tech industry early on are still in the workforce, yet more and more young people are entering the workforce because "learning to code" was the perceived path to prosperity half a decade ago.
> Places want to hire younger people because they're more apt to work longer hours for less pay.
This is the best light you can shine on the discrimination. Most often it really is managers taking their “seniority” literally. As in, they don’t want to take the risk their reports are smarter, more experienced or capable of replacing them, so they discriminate on the basis of age. It’s counterintuitive, but this feels truest from my historical observation.
I wonder if this is really true or just an off-the-cuff comment (no disrespect). I have worked for a number of major tech companies and have never seen this in practice. Some managers were older, some were younger, some the same age. In most/all cases, people were hired based on their technical skills not because they fell into some magical age gap.
> Places want to hire younger people because they're more apt to work longer hours for less pay.
If you take age out of the equation, is there supposed to be something wrong with preferring to hire people who are willing to work longer for less pay over people who aren't willing to work as long who want more pay?
I've experienced the opposite where some smaller companies won't even look at a resume for someone under 30. One of the owners admitted it to me later on. :(
This should also be illegal in the US.
40 and over is a “protected group” in US employment law. 30 year old people are not protected.
It will be fascinating to see the facts of this case, but if it is proven their algorithms are discriminatory, even by accident, I hope workday is held accountable. Making sure your AI doesn't violate obvious discrimination laws should be basic engineering practice, and the courts should help remind people of that.
An AI class that I took decades ago had just a 1 day session on "AI ethics". Somehow despite being short, it was memorable (or maybe because it was short...)
They said ethics demand that any AI that is going to pass judgment on humans must be able to explain its reasoning. An if-then rule says this, or even a statistical correlation between A and B indicates that would be fine. Fundamental fairness requires that if an automated system denies you a loan, a house, or a job, it be able to explain something you can challenge, fix, or at least understand.
LLMs may be able to provide that, but it would have to be carefully built into the system.
I'm sure you could get an LLM to create a plausible sounding justification for every decision? It might not be related to the real reason, but coming up with text isn't the hard part there surely
> I'm sure you could get an LLM to create a plausible sounding justification for every decision.
That's a great point: funny, sad, and true.
My AI class predated LLMs. The implicit assumption was that the explanation had to be correct and verifiable, which may not be achievable with LLMs.
It seems solvable if you treat it as an architecture problem. I've been using LangGraph to force the model to extract and cite evidence before it runs any scoring logic. That creates an audit trail based on the flow rather than just opaque model outputs.
It's not. If you actually look at any chain-of-thought stuff long enough, you'll see instances where what it delivers directly contradicts the "thoughts."
If your AI is *ist in effect but told not to be, it will just manifest as highlighting negative things more often for the people it has bad vibes for. Just like people will do.
Yes, they will, they'll rationalize whatever. This is most obvious w/ transcript editing where you make the LLM 'say' things it wouldn't say and then ask it why.
It sounds like you're saying we should generate more bullshit to justify bullshit.
They said "could", not "should".
I believe the point is that it's much easier to create a plausible justification than an accurate justification. So simply requiring that the system produce some kind of explanation doesn't help, unless there are rigorous controls to make sure it's accurate.
> Fundamental fairness requires that if an automated system denies you a loan, a house, or a job, it be able to explain something you can challenge, fix, or at least understand.
That could get interesting, as most companies will not provide feedback if you are denied employment.
Fair point. Maybe the requirement should be that the automated system provide an explanation that some human could review for fairness and correctness. While who receives the explanation may be a separate question, the drawback of LLMs judging people is that said explanation may not even exist.
I hate this. An explanation is only meaningful if it comes with accountability, knowing why I was denied does me no good if I have no avenue for effective recourse outside of a lawsuit.
This is the law in the EU, I think
the way i understand it is that the law says decisions must be reviewed by a human (and i am guessing should also be overrideable), but this still leaves the question how the review is done and what information the human has to make the review.
Would love to see some of the liability transfer to the companies using Workday too...
As someone over 40, I couldn't help but laugh at the font size on the site.. I guess they know their audience.
The first HN link I visited in a long time where I didn't have to hit ctrl-+ 3 or more times to increase the font size and make it readable.
lol, exactly my response. I have HN itself at 150% (ctrl-+ 3x).
> allegations include that Workday, Inc., through its use of certain Artificial Intelligence (“AI”) features on its job application platform, violated the Age Discrimination in Employment Act (“ADEA”)
I'm interested to see Workday's defense in this case. Will it be "we can't be held liable for our AI", and will it work against a law as "strong" as ADEA?
ADEA says:
> It shall be unlawful for an employment agency to fail or refuse to refer for employment, or otherwise to discriminate against, any individual because of such individual's age, or to classify or refer for employment any individual on the basis of such individual's age.
I don't see any wiggle room for outsourced decision making to remove the responsibility for the outcome.
Hopefully they have something better prepared because "we can't be held liable for the black box we made" doesn't sound like a recipe for success. In Canada, they've already ruled that a website operator is liable for information that their LLM chatbot gives out.
https://www.bbc.com/travel/article/20240222-air-canada-chatb...
Regardless of merit, this seems like appropriate payback for having to make a new workday account for every single company you apply to a job.
Having done a lot of online applications in the last two years, I often just pass up companies that only have their application on Workday.
Workday is without question the most abominable of the online application systems available today.
Surely you haven't seen taleo.
I covered this for my channel yesterday and was surprised how widespread this is. Like most people are likely affected.
As I remember, Workday was found when a founder is in his 60s.
Feel bad for the next guy who wants to sue them but has to settle for workdaycase2 .com
I never liked these "trust me bro we're court authorized, give us all your PII to join the class action" setups on random domains. Makes phishing seem inevitable. Why can't we have a .gov that hosts all these as subdomains?
Yeah, that’s always my first reaction as well.
Oh snap.