The Hilarious Reality: Faking AI in the Startup Industry
The current state of the AI technology industry, particularly within the startup scene, presents a rather amusing, albeit concerning, picture. While the potential impact of AI on various careers is undeniable and potentially disruptive, the reality on the ground for many self-proclaimed “AI companies” is far less advanced than advertised. It turns out, a significant portion of the AI startup industry isn’t quite what it seems.
The Pervasive Misclassification Problem
As far back as March 2019, a survey by British venture capital firm MMC examined 2830 European startups classified as “AI companies.” The findings were illuminating: only 1580 showed actual evidence of working in AI, while the remaining 1250 – roughly 40% – likely had nothing to do with it. This wasn’t just startups exaggerating; financial analysts were also misclassifying companies as AI-related, potentially because AI companies attract higher investment. This highlights a significant issue of hype outweighing reality in the pursuit of funding, a problem likely not confined to Europe.
Beyond Misclassification: Outright Faking AI
Moving beyond companies simply misclassified, a more concerning and surprisingly common situation involves companies that appear well-positioned in the AI world but are, in fact, lying about their technological capabilities. Often, the “AI” powering these ventures turns out to be “Actually Indian workers” or human labor from other regions, performing tasks marketed as automated and intelligent.
Case Study: Builder.ai’s Human-Powered “AI”
A prominent recent example is Builder.ai, a startup valued at 1.5 billion dollars and backed by Microsoft. They promoted a “virtual assistant” named Natasha, supposedly leveraging AI language models to help build software applications. However, in June 2025, following financial troubles, it became public that Builder.ai was not primarily driven by clever technology but by approximately 700 workers in India effectively pretending to be chatbots.
Evidence of this deception was present years prior. In 2019, when the company was known as Engineer.ai, the Wall Street Journal reported that current and former employees claimed the company exaggerated its AI capabilities to attract customers and investors. This long-standing issue ultimately contributed to the company’s bankruptcy, raising the question: how many more companies are doing this?
Amazon’s “Just Walk Out”: Human Oversight Behind the Scenes
Even major corporations have faced scrutiny regarding the true extent of AI in their public-facing technologies. Amazon’s “Just Walk Out” retail technology, marketed as using “advanced AI, sensors, computer vision” for a seamless checkout experience, was reported by “The Information” to be heavily reliant on human workers in India. These individuals would manually review and verify a significant percentage of orders by watching customer activity via cameras.
While the concept was innovative from a customer perspective, the reality was less about cutting-edge machine learning and more about human monitoring. Although Amazon has explored alternatives like “Dash Carts,” which rely on scanner systems, the initial implementation of “Just Walk Out” highlights the use of human labor masked as advanced AI.
The Presto Voice Discrepancy: Media Claims vs. SEC Filings
Presto, a startup providing automated ordering solutions to fast-food chains, offers another striking example of inflated AI claims. In late 2021, the CEO claimed to media that their system handled over 95% of orders without human intervention, touting a massive upcoming national rollout.
However, a November 2023 SEC filing painted a very different picture. Presto Automation stated that over 70% of orders taken by their “Presto Voice” solution *required* human agent intervention, and they were hopeful that the percentage not requiring human help would reach “30% or better” in the future. This drastic difference between public statements to the media and filings under oath to the SEC raises serious questions about potential fraud.
Nate.tech: A Fraud Indictment for Faking AI
The pattern of faking AI capabilities extends to the e-commerce sector. Nate.tech, an AI e-commerce checkout company, advertised a mobile app supposedly using an AI solution to process orders effortlessly. Yet, in April 2025, a fraud indictment in New York became public, alleging that the company’s founder defrauded investors with fabricated AI capabilities while secretly employing personnel to create the illusion of automation.
The indictment detailed how Nate.tech used hundreds of contractors, or “purchasing assistants,” in the Philippines to manually complete purchases via the app. Even when the CEO directed the development of “bots,” the company continued to rely on a combination of these basic bots and low-paid human workers to impersonate an AI program. This company managed to attract $51 million across three funding rounds based on these allegedly fraudulent claims.
Recap: Billions Invested in Illusion and Human Labor
These four examples alone illustrate a significant trend where billions in investment capital have flowed into companies claiming to revolutionize various sectors with AI, only for the reality to involve substantial reliance on underpaid human labor:
- Builder.ai’s app development “AI” was largely Indian developers.
- Amazon’s “Just Walk Out” retail tech utilized overseas workers in India for verification.
- Presto’s fast-food ordering “AI” required human intervention for over 70% of orders, despite earlier claims of 95% automation.
- Nate.tech’s e-commerce checkout “AI” was primarily human workers in the Philippines, leading to a fraud indictment.
It reveals a sometimes-hilarious, sometimes-alarming picture of an industry built on exaggerated claims.
Conclusion: AI Reality vs. Industry Bubble
While acknowledging the genuine potential and future impact of AI automation, the current state of the “AI industry” appears significantly bloated and, in many cases, built on questionable foundations. “Artificial Intelligence” has become a buzzword attracting massive investment, potentially fueling multi-billion dollar fraud schemes and self-enrichment.
The level of hype absorbed by venture capital firms is staggering. Even if the underlying technology promises future upheaval, the current AI “industry” resembles a bubble. It seems unsustainable, with many companies likely facing failure because, as the MMC study suggested, nearly half may not truly be AI companies, and a significant portion of the rest appear to be engaged in outright deception. The reality is that a substantial part of the AI startup landscape today consists of misclassification and outright fraud, perhaps a “40% joke, 30% fraud” scenario, leaving only a small percentage representing actual, robust AI innovation.
