The history of AI is riddled with periods of extreme enthusiasm. Already in the 1960s, for example, the promoters of AI promised AI-powered, autonomous vehicles. The most dramatic wave of enthusiasm started in the early 2010s, when AI became commercially successful and widely adopted for the first time. But as AI made strides, the hype around it reached unbelievable proportions. We’ve seen it all: companies that pretended to use AI when they weren’t, 200 billion dollars invested in self-driving cars that weren’t delivered, a Google employee who told the world their AI was sentient, and so on. And when we thought things may calm down, ChatGPT was released.

In this talk, we’ll discuss the mechanics that stoke the AI hype and their dubious morality. I’ll tell you about shady tactics used by researchers to generate excitement about AI, and we’ll see that inaccurate, sensationalist claims about AI are often published in the highest-caliber scientific journals. We’ll also see how companies abandon solid principles of engineering design and lean growth in order to build cutting-edge AI, and when their efforts fail they often hide the truth from clients and investors. After working in this field for nearly a decade, both in the academic and business AI worlds, I’ve seen some things I wish I hadn’t and I’ll share them with you today.