LinkedIn Should End ‘Diversity in Recruiting’ Feature: “Discrimination by algorithm is still discrimination”
Equal Protection Project calls on LinkedIn to disable the DIR feature which injects race and other protected statuses into the candidate pools promoted by LinkedIn’s algorithms: “Discrimination by algorithm is still discrimination, and LinkedIn needs to stop.”
LinkedIn is one of the most influential platforms for seeking employment and recruiting employees. LinkedIn offers various programs and tools for both candidates and recruiters/employers, with a core promise of non-discrimination in all aspects of the process.
But that’s a promise to which LinkedIn does not live up, instead, embedding discrimination in the platform through the “Diversity in Recruiting” (DIR) feature. DIR, according to LinkedIn’s own descriptions on its website, utilizes demographic information, provided by users who give LinkedIn permission to use such demographic data, to create diverse pools of candidates to present to recruiters/employers. These diverse pools by their very nature differ from whatever other pool of candidates the algorithms would present based on factors such as education, experience, and so on. DIR is intended to add race and other protected categories into what the algorithms pull and present.
The Equal Protection Project (EPP) wrote to LinkedIn in late April 2023, calling on LinkedIn to disable the DIR feature as it violated LinkedIn’s own non-discrimination policies, and also implicate federal and state anti-discrimination laws. We heard nothing back from LinkedIn, and the DIR feature continued to operate.
On July 5, 2023, after the U.S. Supreme Court issued its Opinion in the Harvard/UNC cases, we again wrote to LinkedIn calling on it to disable the DIR feature. The full letter is embedded at the bottom of this post. Here is an excerpt [footnotes omitted]:
We write to follow up on our letter of April 25, 2023, in which we called on LinkedIn Corp. (“LinkedIn”) to disable the “Diversity in Recruiting” (“DIR”) feature on its platform. The DIR feature utilizes the personal demographic information of LinkedIn users to “surface qualified members” in order to “diversify the group of candidates displayed to recruiters …. from companies that have made public commitments to diversity, equity, and inclusion.”1 To date, we have received no response from you, and the DIR feature appears to remain in use by LinkedIn.
We once again call on LinkedIn to disable the DIR feature, particularly in light of the U.S. Supreme Court’s recent decision in Students for Fair Admissions Inc. v. President & Fellows of Harvard College, __U.S.__ (2023) (“Students for Fair Admissions”). In Students for Fair Admissions, the Supreme Court reaffirmed that drawing “[d]istinctions between citizens solely because of their ancestry” is “odious to a free people whose institutions are founded upon the doctrine of equality,” and declared that preferences based on racial classifications constitute invidious discrimination in violation of the Equal Protection Clause of the 14th Amendment, which the Court noted would also constitute a violation of the Civil Rights Act of 1964 (Majority Op. at n.2). It therefore is more imperative than ever that LinkedIn discontinue the use of the DIR feature.
In addition to legal requirements, LinkedIn has made contractual promises of non-discrimination that are consistent with Students for Fair Admissions. Yet your platform contradicts those promises by embedding racial and other protected classifications into the job recruiting process through the use of algorithms that drive the DIR feature. In light of the Supreme Court’s pronouncement in the Students for Fair Admissions cases that “[e]liminating racial discrimination means eliminating all of it,” we again call on LinkedIn to disable the pernicious DIR feature.
* * *
While we are pleased to learn that LinkedIn changed its language [as to employer ability to filter] in response to our prior letter, and that LinkedIn now asserts that potential employers are unable to screen candidates on the basis of protected categories, LinkedIn’s pledge on that score does not go far enough. Assuming the platform now does not allow employers to filter job candidates by protected demographic categories, LinkedIn is doing that filtering for employers, so the result is the same. LinkedIn, by the terms explained on its website, is making candidates more visible to potential employers based on the candidate’s race, ethnicity, gender, gender identity and sexual orientation. This results in a manipulated pool of candidates which discriminates against candidates who do not meet the diversity requirements.
It is no defense that candidates elect to use the DIR feature. The DIR feature affects all candidates – those who opt in potentially get a boost if they meet the diversity requirements, while those who do not opt in obtain no advantage. Either way, diversity candidates are advantaged. The entire recruitment process thus is impacted by the DIR feature for those who participate and those who do not.
* * *
LinkedIn also has not explained what role, if any, employers have in the process. Are recruiters and employers aware that they are receiving a manipulated pool? The website is far from transparent on that point, and its ambiguous language could reasonably be interpreted to suggest that potential employers are made aware that LinkedIn is suppressing “non-diverse” job applicants and “surfacing” the “diverse” ones. While this would be troubling – indeed, as mentioned in our initial letter, it could trigger liability against LinkedIn for violating state and federal civil right laws under an aiding and abetting theory – it would be just as bad if LinkedIn is, in fact, concealing its filtering and manipulation of the candidate pools from the recruiters and employers who are simply looking for the best qualified job applicants. LinkedIn needs to be transparent about its process.
Either way, the DIR feature does not promote fairness or address bias, but itself appears to be engaging in discrimination on the basis of protected categories.
* * *
Information obtained and requested through the pre-employment process should be limited to information essential to determine if a person is qualified for the job – race, ethnicity, gender and sexual orientation are irrelevant to such determinations as a matter of LinkedIn policies and the law. Such discrimination simply cannot be justified as job-related or consistent with business necessity. That was the law prior to Students For Fair Admissions, and if there were the slightest doubt, the Supreme Court once and for all settled the issue. LinkedIn should take notice and adhere.
We call on LinkedIn to live up to its own non-discrimination policies and the law, and to immediately disable the “Diversity in Recruiting” feature. Please confirm LinkedIn’s intention to do so.
LinkedIn has not responded to us as to the July 5, 2023, letter (just like it never responded to our late April 2023 letter), but it did provide a statement to the NY Post which covered our letter, LinkedIn’s race diversity data collection is discriminatory: legal watchdogs:
A conservative legal watchdog group is demanding that LinkedIn take down its “Diversity in Recruiting” feature, claiming it’s in violation of last week’s Supreme Court ruling against race-based affirmative action in higher education.
The Microsoft-owned tech company professes to have 930 million members in more than 200 countries worldwide and acknowledge it collects demographic data that “may diversify the group of candidates displayed to recruiters.”
The Equal Protection Project attorneys, William Jacobson and Ameer Benno, sent their concerns to the LinkedIn headquarters in Sunnyvale, Calif. on July 5 after the nation’s highest court struck down race-conscious admissions programs at Harvard and the University of North Carolina as unconstitutional.
The implications for data used in recruiting is now resonating as a legal danger as employers are reassessing their diversity programs and preparing for possible legal challenges….
In a statement to the Post, LinkedIn defended its diversity efforts for members and job hunters as proper.
“One way to make sure people have equal access is to understand the gender, race, ethnicity, and other important demographic information of our members and then to measure whether or not all are being afforded opportunities equitably on our platform.”
“Many hirers use LinkedIn to connect with diverse groups of qualified candidates for their roles. We often evaluate new ways to ensure that our tools recommend qualified candidates, but we don’t enable recruiters to filter out or exclude applicants based on their race, gender or other protected demographics,” the company said.
“People who use [our site] choose whether to share their demographic information and can control how their data is used through their settings.”
But the lawyers disputed the company’s explanation, saying LinkedIn is the one sorting and filtering members based on race, ethnicity, gender and sexual orientation for the recruiters.
LinkedIn’s defense that some users consent to use of their demographic data was addressed in our letter before LinkedIn issued its statement – consent to use of data is not consent to discrimination, and in any event, consent to a discriminatory system would not still leave the system discriminatory.
I talked about the problems with the DIR feature during an appearance on Varney & Co. on Fox News:
I also was quoted on the “consent” defense and other aspects of the program in an article at Fox News Digital, LinkedIn under fire for ‘Diversity in Recruiting’ feature: ‘Manipulated pool of candidates’:
An advocacy group is pressing LinkedIn after the social media company used their “Diversity in Recruiting” (“DIR”) feature to “diversify the group of candidates displayed to recruiters.”
The Equal Protection Project of the Legal Insurrection Foundation (EPP) is calling for the career social media company to change their policy and not allow a “manipulated pool of candidates.” ….
“The Supreme Court recently held that universities may ‘never use race as a stereotype or negative’ and a ‘student must be treated based on his or her experiences as an individual—not on the basis of race,” Bill Jacobson, president of the EPP, told Fox News Digital.
“LinkedIn’s filtering of demographic data to manipulate pools of applicants to achieve diversity violates these principles,” Jacobson continued….
“Some people are promoted to potential employers based on the candidates’ protected status, which runs counter to LinkedIn’s own non-discrimination rules,” Jacobson said to Fox News Digital. “What’s worse, candidates may never know how they have been treated in this filtering system based on their protected status, and it’s not even clear if recruiters/employers are aware they are receiving a manipulated pool of candidates.” ….
“It is no excuse that some candidates opt-in to having their demographic data used,” Jacobson added. “LinkedIn asking candidates to consent to a discriminatory system does not excuse the discriminatory system.”
The Equal Protection Project alleged that LinkedIn’s algorithm favors diverse candidates-while discriminating against candidates who either choose not to divulge protected status information or candidates who do not meet diversity requirements.
“LinkedIn’s DIR program is particularly pernicious because it builds racial and other protected status classifications into the hiring process through demographic filtering and promotion,” Jacobson said. “This built-in discrimination operates in the background through algorithms at the LinkedIn hub that connects candidates to recruiters/employers to impact the entire LinkedIn-related hiring process.” ….
“Discrimination by algorithm is still discrimination, and LinkedIn needs to stop,” Jacobson said.
A LinkedIn spokesperson told Fox News Digital in a statement, “We believe all of our hiring products and algorithms comply with U.S. laws and regulations relating to discrimination on the basis of protected categories. Many hirers use LinkedIn to connect with diverse groups of qualified candidates for their roles.”
“We often evaluate new ways to ensure that our tools recommend qualified candidates, but we don’t enable recruiters to filter out or exclude applicants based on their race, gender or other protected demographics. People who use LinkedIn choose whether to share their demographic information and can control how their data is used through their settings,” the statement continued.
LinkedIn does not deny what it is doing, and even admits it is servicing a desire from employers for a “diverse” workforce, so LinkedIn appears to be facilitating an employer desire to recruit based on protected categories. That admission sheds a little bit of transparency as to how the program works, something not disclosed on the LinkedIn website.
We will continue to pursue this and will report on new developments.
——————————–
DONATE
Donations tax deductible
to the full extent allowed by law.
Comments
It’s an ‘open and notorious’ act, to borrow a phrase from property law. Perhaps that’s the way to proceed; treat every instance of active discrimination the same way we would a neighbor attempting to steal our property by altering the fence line. If we don’t aggressively push back then the end result is very similar, the adverse possession of our institutions and our liberties.
LinkedIn is for sheep. If you think outside their box you get cut off. I worked for a company that demanded an account so they could keep track of what you said. It became bad enough a story would be posted and every sheep would like it no matter how wrong it was.
Remember: Microsoft owns Linkedin, just like Jeff Bezos owns the Washington Compost.
The truly nasty part about this kind of discrimination is that it puts square pegs in round holes for no good reason (skin color). We have a big country with diverse businesses. Scooping up the wrong person leaves a better fit unfilled.
It is time for a federal divorce.
It’s our only hope of a future of safety and freedom.
Suing Microsoft is the only way to resolve this. The company isn’t going to voluntarily remove this racist and discriminatory feature; that’s obvious.
“Discrimination by algorithm is
stillthe most efficient form of discrimination…”“DIR (Diversity in Recruiting) is intended to add race and other protected categories into what the algorithms pull and present.”
Uh, ok? The plain fact is there are simply too many problem people in protected classes that my small business simply cannot afford to baby sit. Which precisely why I find DIR to be a useful tool that allows me to be a more discriminating employer, more efficiently, by alerting me who not to hire.
Let the hurty-word comments commence!
As I recently saw, one manager says he throws out all resumes with pronouns on them. He figures they are z lawsuit waiting to happen.
Why hire someone who know from the get-go is either a mindless weaking or a nut?
No “hurty-word comments” needed. When someone brags about being a racist, no comments are needed.
“Discrimination by algorithm is still discrimination”
You don’t interact with an algorithm, or machine; you interact with the people who built the thing — they put their choices and preferences into it.
Algorithms are more powerful and insidious than embedded choices, extending the builders’ thinking to examples they didn’t consider directly. An algorithm that works in complex situations is hard. There’s a parlor game these days of posing decision dilemmas to a hypothetical self-driving car.
LinkedIn, meaning the people within LinkedIn, the people behind LinkedIn, the people who use LinkedIn, and the chattering commentarior of regulatory overlords, all have opinions about rights, categories, and “divesity”, baked into the implementation they use.
When do we start suing people for their action by remote control?
“One way to make sure people have equal access is to understand the gender, race, ethnicity, and other important demographic information of our members and then to measure whether or not all are being afforded opportunities equitably on our platform.”
So
1 — Quotas
2 — Bunches of assumptions about how hiring distributions relate to the population mix on LinkedIn. So many pamed cards.
3 — Who understands? Promote themselves to monitors, n compliance-bots, not doing the work. Everybody wants to be a reviewer, not a doer. Way safer.
“in any event, consent to a discriminatory system would not still leave the system discriminatory.”
I think you didn’t mean that “not.”
I celebrate this post.
So, I had been meaning to close my linked in account after learning about Reid Hoffman. So, I did it. It seemed to be easy to do, unlike closing my facebook account in May 2017. The ‘manage account” section contains a ‘close account’ tab at the end of the options list. You’ll need your password. The process asks why you are closing your account. They offer you an option for a customized reason, in which I told them:
“Because democrat donor, LinkedIn founder Reid Hoffman, is funding E. Jean Carroll’s sexual assault and defamation law suit against President Donald J. Trump.
Christine Blasey-Ford
Ms. Judy Munro-Leighton
Deborah Ramirez
All three lied before congress during Kavanaugh’s confirmation hearing. All three got away with it. Ms. Judy Munro-Leighton’s accomplice, then senator Kamela Harris, got away with it too. Women like E. Jean Carroll have destroyed the credibility of America’s genuine victims. Shame on you. You fell for the democrat lies.”
Feel free to modify this if you choose to dump the liars. Y’all have a nice Summer.
FJB.
Trump 2024.
none of this racial nonsense will change until all “protected categories” are abolished. That’s the real issue.