The founder of an app that supposedly uses AI to help with online shopping is accused of fraud. In reality, the vast majority of the supposedly automated transactions were carried out by humans.
Have you heard of the Mechanical Turk? It's a chess-playing automaton created at the end of the 18th century. For over 80 years, it was believed that it was actually a robot capable of moving the pieces on the chessboard and winning on its own. In reality, a human was hidden in the large table used for the supposed mechanics of the machine and controlled it like a puppeteer. The story of Albert Saniger is quite similar, except that the automaton is replaced by artificial intelligence and the chess player by call center employees.
Read also – This new artificial intelligence wants to revolutionize scientific research by making its results easier to understand
In 2018, the man from Barcelona, Spain, founded Nate, a company offering an app presented as revolutionizing online shopping using AI. Its promise: to save you time by taking care of all the steps that follow the spotting of an item on a sales site.
Example: you open the app, find a pair of shoes you like and press the “Buy” button. The AI then takes care of choosing the right size, entering the delivery and payment information, and validating the transaction. Convenient, right? Except that under the hood, that’s not really what’s happening.
The creator of an AI app is accused of fraud, humans were doing the work
When seeking to raise funds from investors, Saniger presents his proprietary AI system as “capable of carrying out online transactions without human intervention”. Aside from a few “extreme cases”, everything is automatic. In reality, although he purchased third-party AI technology and hired a team to develop it, it doesn't work at all. While the man is touting his product, no transactions are being carried out by artificial intelligence.
Read also – He creates an AI tool to land a job at Facebook, Amazon, or TikTok, the result isn't quite what he expected
Saniger secretly hired hundreds of real-life employees, including those from a call center in the Philippines, to handle transactions made via the app. In the fall of 2021, when the online shopping season was approaching, he asked his teams to create bots to support the workers. Yet he has always said that his system doesn't use them.
At the same time, the entrepreneur does everything he can to avoid suspicion. He orders his employees to remain silent about the rate of transactions actually managed by AI and severely restricts access to the dashboard that records them. When asked why he is so secretive, he replies that the data relating to his system is a “trade secret“.
For having sold AI that did not exist, the entrepreneur risks a lot
It was in 2022 that the media The Information revealed the truth. An investigation was opened and we learned, for example, that in 2021, three years after its launch, “the share of transactions processed by [the application] Nate manually rather than automatically was between 60 and 100%“. In total, Saniger raised $40 million from investors by selling them hot air.
Read also – Artificial intelligence can lie now, this study shows how
Following the FBI investigation, Albert Saniger is accused of stock fraud and wire fraud by the State of New York. For each count, he faces up to 20 years in prison, or 40 years behind bars in the worst case scenario, as American sentences are cumulative. These are the maximum penalties that can be imposed. If convicted at trial, the length of imprisonment will be set by the judge.
Source: Department of Justice
0 Comments