The future of Humans: Operators of AI
2023 has been a year of transformational change in AI. But it has been most transformational for regular humans getting to interact with near-human-level AI for the first time this year. Personally, I got to live that moment five years ago that so many got the chance to experience just recently. For many, awe and anxiety combined in this moment, at once laying bare a preview of our world to come.
In such a short timespan, we’ve seen AI being adopted everywhere, from image generation to text generation. Decision makers that had no idea what a language model was were now exploring how to integrate it with their offerings. Regular folks using it for anything and everything. We’re in the process of picking all the low hanging fruits. And the race for integrating AI means there are a lot of decision makers and leaders making decisions about AI with very superficial knowledge and consideration.
I believe the largest risk in AI is human mismanagement. At least, that’s what I’ve observed in my career. Humans misjudging, misunderstanding, misapplying their capabilities. Even AI professionals, those with postdocs in ML make such misjudgements. The human brain deals in human logic which has its characteristic human biases. Just like it took us a while to figure out what fallacies even were and how not to fall for them, we’ll need to first define AI fallacies and human fallacies towards AI and then teach the public these condensed learnings so that we can easily identify 90% of bad uses of AI before they happen.
A simple prediction for the next few years and decades to come is that many if not most jobs will be gradually augmented or outright replaced by AI. It’s true that we are going to be faced with an enormous disruption in the workforce and this is why I am writing this blog post. I want to make the case for humans as operators of AI, not as victims of AI.
It’s true some people will have their life changed irrevocably by AI. I want to believe that it will help most people go on to better opportunities, but there will be some losers in this equation. I want there to be the least amount of net negative for humans, which is why I am advocating here for the rapid education of all our workforce. Even though the AI may arbitrarily affect workers, the choice is still yours to be that person who welcomes it and leverages it. And you can’t do that if you don’t understand what you’re working with and what’s coming.
I am not envisioning everybody becoming an ML engineer like me and learning how to code, I am envisioning humans becoming operators of AI. Just like the machine or computer or software you operate every single day in your job, you’ll need to learn this new tool. Not only that, you’ll need to understand it well enough that you can be the one managing it. There are no established best practices yet on how to successfully operate AIs. You’ll need to be one of the firsts to figure that out if you want to succeed in our new world to come.
A large portion of your work or team may be replaced by AI. You are given the choice at this moment to get a head start and be the one that multiplies their effort instead of being the one that suffers from it and lets others suffer as well. If you understand AI well enough, you’ll find opportunities for it to help yourself and others. Get your work done faster and better. You may even get to go through your entire backlog of work that’s been sitting there, waiting to be picked up forever.
I want the public to learn from my experience in AI to hopefully catch bad outcomes before they happen. I want our use of AI to be responsible, so that risk can be managed. Hopefully, humanity gets to continue its progress, maximizing positive outcomes for all.
On this blog, I will be writing with the aim of making the public better at understanding their interactions with AI, better at using it for their own purposes.
As a follow-up to this post, I encourage you to read The AI Operator’s Handbook.
