The Digital Divide

 The Digital Divide has been a concern of policy-makers and observers of American society since the late 1990s. Many today view the issue as being in general solved; the vast majority of people today have a smartphone, access to a computer, and the internet.  Indeed, great strides have been made to bridge the gap and connect all Americans with the information and opportunity available online.  Factors like the availability of Wi-Fi in many public places, the widespread adoption of technology sharing programs by school districts across the country, and the real price deflation of many tech products have made the internet and worldwide connection more realistic to all Americans than ever before.  For much of the previous ten years, the perception mentioned above that the digital divide was practically a solved problem was broadly correct, in my opinion at least.  Americans certainly had disparate levels of understanding of how to best use the internet, but almost everyone had access.  Artificial Intelligence has changed that; access to the newest models is only available behind a paywall, which few opt into. Furthermore, despite the raw popularity of the technology, the lack of sophistication that the majority brings to their engagement with AI makes it a difference not in degree but in kind from previous disparities in skill.  Power users of AI report huge productivity improvements, especially from those in technical fields. The average user, in comparison, treats AI more like super Google, or as a plaything. In summation, the huge difference between the popularity of Large Language Models and their hypothetical efficient use case is a modern digital divide, strongly akin to the physical one of the turn of the millennium.