NEW YORK — The shopping startup Nate, Inc. told venture investors its proprietary artificial intelligence automated 93 to 97 percent of online checkouts. According to federal prosecutors and the S.E.C., the actual automation rate was closer to zero — unless you count human contractors typing orders by hand in Manila and Romania.
On April 9, 2025, prosecutors in the Southern District of New York indicted founder and former chief executive Albert Saniger for securities and wire fraud, while the S.E.C. filed a parallel civil complaint. Both filings allege Mr. Saniger raised more than $42 million by selling Silicon Valley on a breakthrough neural network that did not work. His criminal trial is proceeding in 2026; he is presumed innocent unless proven guilty.
The Mechanical Turk Pitch Deck
Nate’s pitch was built for a frothy fundraising cycle: a universal checkout app that used machine learning to complete purchases across any e-commerce site without human intervention. Investor presentations touted 93 to 97 percent autonomous completion rates, helping bring in more than $42 million.
Behind the sleek loading animation, filings allege, the “neural network” was an overnight shift. Whenever shoppers tapped buy, orders were routed to human workers in the Philippines and Romania who manually keyed payment and shipping fields into merchant websites.
Regulators have taken to calling the genre “A.I.-washing.” If prosecutors prove their case at trial, Nate will stand as its purest artifact: $42 million in venture capital deployed to discover that the cheapest way to fake artificial intelligence is still human intelligence on the graveyard shift.