AI Talk: AI is hard and “Futureproof”

May 7, 2021 / By V. “Juggy” Jagannathan, PhD

This week I explore two polar opposite views: First, AI is hard and is nowhere close to performing at a human level, and second, AI is going to take away our jobs, so we must protect ourselves from being considered surplus.

Why AI is Hard

I came across a YouTube video, which helped explain a paper written by Melanie Mitchell titled: “Why AI is Harder Than We Think.” Both the paper and the YouTube video explanation (with a different point of view) were interesting to read and to watch.

The basic thrust of the paper: AI has seen cycles of springs and winters ever since its inception in 1950s. Hype surrounding Perceptron and expert systems of the 1980s led to a period of disillusionment with AI, dubbed the “AI Winter.” Deep learning is riding high now, but the author warns that the output of current day systems cannot really be called human-level understanding. She highlights fallacies in thinking about AI systems. Just because AI systems can solve a problem in a narrow domain doesn’t imply a more general ability. What is easy for humans is hard for machines (like commonsense and perception) and what is hard for humans is easy for machines (like computation). Whether or not we will see the type of disillusionment that the field faced in late 80s is questionable; however, the promise of driverless cars and other predictions about AI capabilities have yet to be fully realized.

Futureproof – By Kevin Roose

Kevin Roose is a New York Times technology columnist and has been writing and blogging about tech for over a decade. His new book Futureproof: 9 Rules for Humans in the Age of Automation is all about dealing with the impact of AI and automation. There is a relentless march towards automation and the book’s message is that no one is safe. So, how does one survive in such a world? He describes nine rules to help navigate this changing world. All you need to do is look at the book cover to see what those rules are!

An overarching theme in the book is to focus on your humanity and special talents, i.e., what makes you human. Focus on honing your empathetic and social skills and any special talent you have – artists, artisans, athletes and kindergarten teachers have less to worry about in this new world! Humanities, (instead of or in addition to STEM disciplines) will provide a strong underpinning to help us navigate the world of automation.

Don’t become addicted to your devices, taking cues and letting an algorithm control your life. He admits that he himself became addicted to the dopamine rush of constantly monitoring his likes on Facebook, Twitter and Instagram (I feel lucky that I don’t use any of these social media outlets). Roose uses the term “Attention Guarding” to refer to our ability to focus and get things done without being distracted by external influences. One of my favorite activities gets a thumbs up – meditation and afternoon naps!

The author urges caution on the use of “bureaucratic bots” that make life changing decisions, from helping with job recruitment to approving loans and deciding who qualifies for various social safety net services. These bots have repeatedly been shown to be biased, causing more harm than good. Treat bots as you would use a “chimp army,” meaning never or with very, very careful planning and supervision.

He underscores the importance of revamping our social safety net and recommends that we reengage on discussions surrounding Universal Basic Income (UBI) and Medicare for All. For an excellent synopsis of the book, listen to the author’s interview on NPR.

Commentary

The first example above says AI capabilities are overrated, while the second example on Futureproof warns of massive disruptions caused by AI. I believe the truth lies between these extremes – at least for this decade. There may be significant job displacement, but blue collar workers could come out ahead. If a massive infrastructure bill comes to pass, there is going to be a big shortage of electricians, plumbers, brick layers, construction workers and handymen. These jobs are definitely futureproof for the next few decades. Nurses and elder care workers will also be in high demand. White collar workers better heed the message from Futureproof, as more and more office work becomes automated! As Roose points out in his NPR interview “..artificial intelligence is not just replacing sort of repetitive manual labor. It’s also replacing repetitive cognitive labor.”

I am always looking for feedback and if you would like me to cover a story, please let me know! Leave me a comment below or ask a question on my blogger profile page.

V. “Juggy” Jagannathan, PhD, is Director of Research for 3M M*Modal and is an AI Evangelist with four decades of experience in AI and Computer Science research.

Listen to Juggy Jagannathan discuss AI on the ACDIS podcast.