Artificial Intelligence: Human UnderstandingArtificial Intelligence: Human Understanding This is the mini-conclusion to our mini-series on Artificial Intelligence! Last time we spoke about the practicalily of these tools. We talk about human understanding and finish with the social/economical impact these tools will have. We break out our big brush as we paint broad strokes. On a more meta note, I see "move fast and break things" applies directly to my blogging habits and my treatment of the English language, respectively. As the limits of computation and storage are pushed out, the less we can rely on human understanding to produce results or verify the work of machines. Machines are already performing work that humans cannot easily understand or verify. Human Understanding The "Artificial Intelligence" series resumes with part three - Human Understanding. In a number of tasks, computers surpass human capability. I think people can generally agree with me on this, but it doesn't often get spelled out in a practical manner, or a manner that is easy to understand. Trusting Computer Output We borrow the example from our previous discussion on the practicality of Artificial Intelligence. If a computer gives you the first 10 million digits of PI - do you trust the result, and how quickly could you verify the output? By our (very rough) calculations, a human lives for about 2.3 billion seconds. In perfect conditions a human could calculate the 2.3rd billionth digit by hand if they started the second they were born (and they didn't make a single mistake.) Since a computer can calculate many more digits in a shorter time, there is a point where computers can accomplish more than a human can in their entire life. Point being, we have to trust the results we get in certain situations (fun fact - we already do, but maybe only implicitly.) We take a look at a couple other examples and talk a bit about the potential of "AI" driven systems. Intelligent Systems Intelligent systems are being relied upon to play substantial roles in our lives. Some are scarier than others (algorithms used to estimate criminal recidivism versus how much water a plant should get) but at the end of the day we will be seeing more and more tasks and processes accomplished by machines. We need to have a solid understand on not only the role they play, but how they will affect our lives. Level 5 Self Driving Cars You've likely seen or heard about autonomous vehicles. They're still a ways off, but they'll be here eventually. As we commute everyday, the continuing rise of self-driving cars is driven by artificial intelligence. These AI applications will be/are relied upon by human passengers - even if fully driverless cars are still on the horizon. Even now, lane control and braking assistance functions are trickling into consumer vehicles. There are interesting ethical considerations with respect to self driving cars (Marvel style setup to future blog post!) Processes and Applications Any kind of a process involving an application (loans, university, jobs) will be increasingly automated. There are also a number of ethical concerns here, but these systems already exist - it is just a matter of time before insurance companies are blaming defrauding their customers because of a "rogue algorithm." Furthermore, these kinds of systems are subject to reflecting institutional bias and designer prejudice just as much as they purport to be magical formulae of objectivity. Intelligent Systems: Economic Consolidation Regardless of application, we are going to see an immense economic consolidation. AI/Machine Learning/Whatever you want to call it is a force multiplier for the people and organizations with the means to deploy the tools. People asking themselves "why would I need to know Calculus II?" will be on the outside looking in as the economy rapidly shifts towards an even higher degree of balkanization between specialists and general labour. Our philosophy behind such advances is to increase human potential and expand capabilities. When we roll out solutions, our goal is to grow with the organizations we choose as strategic partners. Being data-driven, or technologically capable does not mean the need to replace people or reduce the workforce. Another aspect is the reach these tools have. From the automotive industry to shipping and logistics - there is almost nothing that cannot be affected by the new powerful systems built upon the IT infrastructure built in the last 20 years. This is on the scale of the industrial revolution, make no bones about it. That's it for now for the AI mini-series! Very topical, but we want to focus more on the potential issues versus the actual history and math behind these techniques. This is a topic we'll absolutely revisit, but for now Mission Accomplished!
Tap or click on these posts to navigate to the next or previous posts.
This post is part of a larger series. Tap or click on a post to view more in this series.