The six waves of AI in the 21st century

AI as a technology is more than 70 years old and has gone through ups and downs. The downs have been called AI winters and there have been a few of them. Until the 21st century, AI was used mainly in niche areas: Jim Simmons is thought to have been a pioneer investor using Machine Learning to guide his Medallion fund, Ray Kurzweil founded the company Scansoft, which used  Object Character Recognition technology to digitize written text and resulted in the company Scansoft. Aside from such niche successes, AI had little impact in the 20th century. That changed in the new millennium. 

AI has still had its ups and downs, to be sure, but the dynamic is more similar to interacting and oscillating movement in time.  Rather than talk about seasons it seems more appropriate to talk about AI waves since they all follow a similar pattern of a period where the technology is evolving and then a crescendo of high awareness and expectations after which it falls back to obscurity. These waves tend to elevate a few champions and have a lasting effect that can sometimes be more subtle. Some waves have left a bigger impact than others but all focus on different applications of AI. The six major waves we can distinguish in the 20th century are:  

• Search – was the first wave that kicked off the modern quest for artificial intelligence fuelled by the exponential growth of data on the internet 
• Recommendations – were driven by the success of online retailers and the need to tailor sales
• Computer vision – only took off after publicly shared training data became available to provide benchmarks for comparison of algorithms
• Game playing – has been a major driver for conceptualizing what it means for a machine to be human-like
• Voice recognition – was the stuff of science fiction, pure magic that one could imagine talking to a machine
• Generative AI – suddenly created a genuine feeling of conversation with and creative potential in the machine

Some waves spawned multibillion-dollar companies and industries while others fizzled out into the background; some were driven by the availability of new data while others were driven primarily by the imagination. Having been part of the ride for all of these waves it is interesting to see the same optimism every time at the peak that AI is fundamentally going to change everything and to see at the bottom that it has been forgotten but the effects of the technology are ubiquitous. 

Looking at the history and dynamics will help us appreciate the value and limitations of AI. It might also point us toward a better understanding of what drives the advance of AI. Stay tuned over the next few weeks for a deeper dive into this fascinating history of a technology that may fundamentally transform parts of how we live and interact with technology.