Is it time to stop hyping the idea that AI has been hyped?

It has become fashionable to capitalize on the mistakes of individuals, who as humans are quick to anthropomorphize things, who encounter text-based (e.g. GPT-3) AI systems and conclude that they might be sentient. Come on people! Isn't it time to stop hyping the idea that AI has been hyped?

Here we have yet another story, yet another piece in social media, this one in LinkedIn, about the need to stop hyping AI.

Ever since the dawn of the Industrial Revolution in the 1700s, coming shortly after the full flower of The Enlightenment, we have experienced significant, continual profound changes in technology and society. This has led to both people anticipating more change, and to people failing to fully understand and keep up with changes that have already occurred. This includes the mundane, such as user interface changes to browsers and operating systems that wipe out personal productivity gained from daily practice, to profound societal challenges, such as erosion of community cohesion as religious institutions have declined in popularity and not been replaced with equivalent neighborhood-building systems.

With artificial intelligence, even the basic curve fitting model of deep learning that is in widespread practice is both taken for granted and yet not fully understood. When it comes to AGI, it is not the case that most people believe that it exists. Those who believe in dualism basically find the idea of AGI impossible. Then there are AI researchers who, having failed many ways in the past to create AI beyond curve-fitting or Eliza-era symbol manipulation, think that it is impossible for anyone else to create it either. Hubris never so clearly indicated that a contrary event was just around the corner.

The deeply learned cynics have certainly been keen to smash AI anticipation based on Terminator movie -like fears or suspicions. Social media, as always, is prone to promote outrage and the barely believable. Yet is there a sense that people are committing, en masse, to the firm belief that AGI exists right now? I don't see it. The existence of a single blog post, tweet, or Facebook message doesn't prove that there is mass hyping of AGI.

Becoming accustomed to new things involves so many different waves, all overlapping. First there is creation of the thing. It's not always a binary, "it was impossible, now it's practice" event. When the Wright brothers flew their airplane at Kitty Hawk, it was a short flight at low altitude with barely any payload. It took much more experimentation, engineering, and practice to get to the primitive airplanes that fought in World War 1. The same is true of AI. First we have curve fitting, then really good curve fitting, then profound curve fitting. Perhaps then we get some ambiguous results. If the first AGI has an IQ of 40, would you call it intelligent? No, you would call it stupid! It's sentience will be highly questionable. And yet there may never be a moment when you can say, "before this all was not sentient, and after this artificial sentience exists".

The second wave is getting the news out. The researcher does the experiment. The knowledge produced may have to be replicated, shared with other researchers, discussed, written up, then published. This takes months or even years. 

The third wave is belief among many. When the news comes out, did you read about it? Did you get the right message? After all, journalists are abysmal at getting the facts straight in fields they don't experience or understand fully. As it is discussed, some believe it, others discount the news story.

The fourth wave is belief within yourself. You've heard the news, but do you believe for yourself what was conveyed. You have to not just hear it, but have a mental model of it that works. It may take some time to get enough information that you can form an accurate belief.

The fifth wave is understanding the implications of your beliefs. It is one thing to understand a phenomenon. But what does that mean for your profession? For your home life? For your religion, your beliefs, your politics? Does it change your values, how you approach the goals of your life? Does it pose a danger to any of the aforementioned?

The current state of AI and AGI is subject to all of those waves. So articles about "hype" are troughs of such waves. They could be an indication of proper sober reflection. The "experts" still put AGI out five to 15 years or more. That some laypeople are confused to think that it is happening now is well within the adoption wave system.