It was one of those Menlo Park dinners where everyone pretends they’re just decompressing from the week but somehow ends up pitching their latest moral framework. Mr. X had gathered the usual suspects—founders, VCs, and a handful of “ethics consultants” whose job descriptions were as vague as their LinkedIn endorsements.
Eric, who’d recently pivoted his HR startup to “AI culture analytics,” was the first to brag. “We just rolled out our new AI workplace companion,” he said, swirling his pinot. “It’s like a digital HR partner—monitors sentiment in Slack, detects burnout, even schedules what we call chemistry check-ins to maintain team cohesion.”
“Chemistry check-ins?” I asked.
He nodded, proud. “The algorithm pairs employees who have complementary personality vectors. Sometimes that’s a manager and a report. Totally bias-neutral. It’s science.”
Across the table, a startup lawyer choked on his amuse-bouche. “Didn’t you just re-create Tinder for power dynamics?”
Eric grinned. “That’s the beauty of it—it’s automated consent. The system only schedules a meeting if both parties show positive sentiment toward each other. It’s literally impossible for harassment to occur.”
The table laughed, nervously. No one wanted to be the first to point out that the last time Silicon Valley declared something “impossible,” it was cryptocurrency fraud.
Mr. X leaned back, always the provocateur. “So, if your AI sends me to lunch with an intern, and the intern thinks it’s a mentorship opportunity, who’s responsible when TechCrunch calls it a scandal?”
Eric shrugged. “The machine doesn’t have intent. You can’t sue an algorithm.”
Later, over cocktails at the Rosewood, the debate metastasized. Someone compared it to self-driving cars—who’s at fault when ethics becomes automated? A venture capitalist claimed this was the dawn of “bias-free desire modeling.” Another said the real risk wasn’t harassment; it was the AI learning everyone’s preferences too well.
“Imagine,” he said, “an algorithm that knows which compliment you’ll fall for before you hear it.”
Someone joked that the next update would probably include “consent NFTs.” Everyone laughed again, relieved to be in on the irony, but nobody seemed eager to question the underlying faith: that morality, like everything else, could be A/B tested.
By dessert, Eric was defending his product as “a corrective to human error.” Mr. X raised his glass. “That’s the Valley motto right there—why confront a problem when you can just code around it?”
The laughter this time was thinner. Somewhere in the distance, the system was already scheduling their next chemistry check-in.
Leave a Reply