The Dark Side of AI in Hiring: That HigherView Scandal and Why Algorithms Are Sketchy

Ever scratch your head wondering who actually looks at your job applications? Online job hunting, total wild west. One spot, thousands of resumes. Companies promised a fast lane. AI in hiring, bam. Sounded super efficient. Like some slick Silicon Valley trick to cut out the noise. For a while, it was all about making things better. But what if that ‘better’ cost way too much? Turned us all into just numbers.

AI in Hiring: HigherView and the Messy Truth

So, okay. Imagine 2019. Could be anywhere. A young pro gets ready for the job interview. And another thing: No human. Just a cold screen. Blinking camera. Automated questions. Smile. Articulate. Your best. But most totally missed it: Software wasn’t just listening; AI was analyzing. Deciding everything. Yep, welcome to HigherView’s digital interview room.

The HR problem? Huge. The internet blew up applications. Recruiters swamped with thousands, even tens of thousands. These HigherView folks? They jumped in. Sounded like saviors. Simple promise: Not just scan resumes. Actually look at candidates themselves. Find the perfect fit, zip, bang, boom. Plus, supposedly ditch human bias—gender, race, age. Big dreams, right?

HigherView started small. Totally humble. 2004, Mark Newman, some international gig. He hated those dumb, long, expensive phone interviews for overseas people. So he grabbed Chip Luman and Mike Walls. Basic idea. A platform. Record video answers. Recruiters watch ’em whenever. Easy peasy. Just a tool to make logistics work, not some kind of judge.

But then, 2010s hit. Silicon Valley? They just had to optimize everything. Those video hours? A total goldmine. Original idea? So last year. The real revolution, they thought? Not just watching. Mathematically understanding. In 2013, they went hard on AI. Changed everything. Suddenly a tech titan, measuring human behavior. Flush with $45 million by 2015. And they weren’t just saving time anymore. Their new pitch? “Uncover hidden talent. Kill human bias. Totally data-driven, perfect hiring.” CEO Kevin Parker called it HR’s holy grail. Faster. Cheaper. Supposedly fairer.

Emotional AI in Hiring: Bad Idea

So, central to all this tech, some sketchy stuff: emotional AI. This software. It scored you. Tiny face movements. Eye darting. Voice sound. Your very words. Were you excited? Focused? Stressed? Suitable? HigherView said its programs, trained on star employees, could predict job fit. By ‘reading’ digital body language. For companies, this saved time and money. Candidates? Career derailed. Cultural stuff. Bad Tuesdays.

Process went from people talking to machines gathering data. Millions found themselves alone, on digital stages. Trying to impress code. Nobody knew what was inside that “black box.” What expression was a nope? What word was a warning? Supposedly, speed. But this questioned human depth, what’s fair. Before things blew up, Goldman Sachs, Unilever, Hilton. All in.

HigherView’s setup had two main bits: game tests for thinking skills. And another thing: The really crazy one? Video analysis. Their pitch? Real slick: “Give us videos of your best workers. Our AI spots their common ways—how fast they talk. Word choices. Facial expressions. Then, it uses this model on new people. Finds those who look like your top performers.” In theory, cool. In practice? Just copied old biases. Made a homogenous group. Automating unfairness.

Video landed. Super detailed scan. Nobody saw it. This was the “black box.” Inner workings? Total mystery, secret sauce. But leaks, tech experts, even HigherView’s own marketing… they showed a creepy picture of this digital judge. Three layers of filters it used:

First, word checks. Speech to text. Keywords. Sentence build. Active or passive voice. Emotional vibe. Saying “I took responsibility” might score better than “A task was given to me.” This one? Easiest to get. Least wild.

Second, the tougher one: voice checks. Not so much what you said. More about how you said it. Pitch. Speed. Pauses. How you go up and down. All measured. Supposedly showed excitement, confidence, stress. But this really messed with people from chill cultures. Accents. Non-native speakers. A calm voice? Could mean you didn’t care. Seriously unfair.

Third, most messed up: face checks. HigherView’s system zoomed in. Tens of thousands of face points. Micro-expressions. Raised eyebrows. Lip tweaks. Based on a shaky 70s theory. Paul Ekman’s idea. Universal emotions across all humans with the same faces. Scientists called it way too simple. Missed all the context. Critics said HigherView was doing high-tech fortune-telling with faces. Reading character from features. Nervous frown vs. concentration? Who knew? That question. Unanswered.

Black Box Bias: Just Copying Old Problems

Biggest danger? ‘Garbage in, garbage out’. If ‘success’ meant certain schools, accents, social classes, the AI copied that. Exactly. So, didn’t clean up bias. Amplified old prejudices. Huge scale. And it stuck. Mathematical certainty. It was a total digital “black box.” Exactly how it worked? A deep secret. Impossible to check independently. Academics? Said it was all fake. How could faces show competence or honesty?

Public Yelling Forces HigherView to Dump Face Scans

Years pass. AI quietly filtered millions. Rejected folks? Totally confused. Rejection sucks. This just felt worse. On Reddit, stories poured out. Weird. Dehumanizing. “Had to smile like a psychopath the whole time!” “Poster on my wall got me rejected?” Started small. Little grumbles. But they were tremors. Building into a giant wave of anger.

Meanwhile, academics and groups working in tech ethics started raising alarms. NYU’s AI Now warned. Experts said HigherView’s tech was bogus. Faces couldn’t show competence. Total lie.

And then, 2019. Whispers turned fierce shouts. EPIC – the Electronic Privacy Information Center, big digital rights group – filed a formal complaint. Against HigherView. Federal Trade Commission (FTC), watch out. EPIC said HigherView was doing shady business. Using unproven, secret, probably discriminatory tech to size up people. First real legal crunch on their “trade secret” shield. Scandal, ignited.

Media jumped on it. Fast. Washington Post dropped a bomb. Drew Harwell’s article. That title, man: “An algorithm that scans your face is increasingly deciding if you get the job.” It showed the ugly reality. Millions of job applications. How the tech worked. How flimsy its science. And the total helplessness candidates felt. HigherView wasn’t just some work software anymore. Symbol of crazy tech power. Algorithmic unfairness.

Late 2019, early 2020. HigherView was under siege. Fire, started by EPIC, fanned by Post. Unstoppable. Company name? Now stuck with biased AI, creepy hiring, fake science. Big clients freaked out. Their brands! They started demanding answers. Those thick walls around their algorithm were taking hits. Everywhere.

HigherView’s first move? Totally expected. Fought back hard. Blog posts. Statements. Said everyone misunderstood their tech. Swore their real goal was less human bias. Algorithms checked job stuff, not emotions. Insisted face analysis was tiny. Said a human always made the call. But nobody bought it. Not public. Not critics. Refused to peek behind the “trade secrets” curtain. Defenses? Sounded fake. Black box stayed shut. Everyone just imagined what monster lurked inside.

Pressure built. Cracks. Independent auditors. AI ethics experts. Asked to see the algorithm. Check the claims. Mostly ignored ’em. No answers. Just more suspicion. Media grilled companies using HigherView, like: “Seriously? Using unproven tech to filter candidates?” PR nightmare. Problem wasn’t just HigherView’s anymore. Belonged to everyone using it.

And another thing: Then BAM, January 2021. Nobody saw it coming. HigherView said: facial analysis. Gone. That main part they’d fought for years? Poof. Scandals peak. They tried to say it was about ethical AI. But from outside, it was a retreat. Crushing public pressure. Dumped the most controversial part of their “scientific” tech overnight. Critics? Saw them admit failure. Indirectly, anyway.

Face Scans Gone, But Other Biases Lurk

But the sneaky bit? That was the thing. Facial analysis? Gone. Yet, systems checking voice tone and word choice? Fully on. They dumped the obvious one. Kept the sneaky stuff. The problematic core. Not knocking down the whole wall. Removing the weakest brick. Because, the thought of companies just making silent, effective filtering totally normal? Super concerning. Still just a score.

Regulations Hit Hard: NYC Law 144 Leads The Way

Ditching face analysis in Jan 2021? HigherView’s stopgap. Didn’t work. Looked like weakness. Admits they were wrong. Fired up lawmakers. Wild west? Over. The fight moved from courts to actual laws.

Most concrete thing after this struggle? First serious AI oversight regs in the US. Globally, too. New York City? Led the charge. July 2023, NYC Local Law 144. Revolutionary. Companies in New York using automated tools (AEDTs) for hiring/promotion? Gotta publish annual, independent audit reports. Prove no race or gender discrimination. And candidates? Had a right to know a machine was checking them. Plus details about the data. The black box? Finally cracking open. By law.

Similarly, Illinois passed an AI Video Interview Act, too. Candidates got rights. Wanted to know exactly how companies used video analysis. On the federal side, EPIC’s complaint didn’t get a direct fine. But it totally changed the FTC’s position on AI bias. From 2021, they said: selling or using algorithms that cause discrimination? That’s unfair dealings. Subject to investigation. Existing consumer protection laws. Crystal-clear message to all of tech: you can’t hide behind your algorithm. You are responsible for its outcomes.

New AI Reality: Ethics vs. Efficiency

HigherView survived financially. Yeah. Rebranded. Talked more about being open, ethical. Lost some clients. But mostly kept their share. But boy, the scandal rattled all of HR tech. Now, big companies buying AI? Not just about fast and cheap. Gotta check legal risks. Audit reports. Ethics. Bias auditing? Whole new business.

Societal cost, though? Deeper. AI got real. Directly impacted lives. Unsupervised power. Pulled back the curtain for millions. Trust in tech? Gone. Human dignity forgotten. Just for speed.

So, where are we now? Years post-scandal. Landscape divided. One side: good stuff happened. More awareness. New laws. Huge pressure on companies for AI ethics. HigherView pulling back from face analysis? A big deal moment. Proving even tech giants can be reined in. But on the flip side? Sneaky danger. Normalization. HigherView dumped the creepy-looking features. Quieted the crowd. But it kept the guts. The system judging you by voice. By word choice. The real risk? By getting rid of the obvious bad methods, companies might make less visible, but just as problematic, algorithmic filtering an everyday thing. Face scanning scandal? Might just be a sacrifice. Paving the way for more subtle robotic gatekeepers.

Right now, owning our own data. That feels crucial. And another thing: A job interview video? Not just simple answers. Your biometrics. Your chat style. Your whole vibe. Who owns that? Store it forever? Train other AI later? When efficiency wins out over fairness, systems lose their soul. They stopped reading faces. But still see you as a score. Just data.

Frequently Asked Questions

Q: What was HigherView’s “emotional AI” thing?
A: It was tech. Looked at tiny face stuff. Eyes. Voice sound. Your words. Claimed it could guess if you were excited, focused, stressed. Predicting job fit. Based on some pretty shaky psychology.

Q: Why did HigherView get rid of the face analysis?
A: Lots of reasons. University critics, like NYU, spoke up. EPIC filed a big complaint with the FTC. Media blew it up—Washington Post, huge article. Everyone got mad! Public outrage. Forced HigherView to ditch the controversial face scanning in January 2021.

Q: New rules after the AI hiring mess?
A: Yep. HigherView scandal led to big ones. NYC Local Law 144, awesome. Now companies using automated employment decision tools must post audit reports. Prove no discrimination. Tell candidates a machine is checking them. Illinois has an AI Video Interview Act. FTC? Says discriminatory algorithms break the law. Big changes.