View profile

🤖 [CITY.AI] - Applied Artificial Intelligence vol.3 (By Azeem Azhar)

City AI
🤖 [CITY.AI] - Applied Artificial Intelligence vol.3 (By Azeem Azhar)
By City AI • Issue #31 • View online
🔥 We'r thrilled to announce that our third edition is curated by Azeem Azhar, curator of Exponential View and experienced tech & AI investor. Furthermore, we have discounts for TechCrunch Disrupt Berlin on 4-5 December - scroll down for more info.
World Summit AI is happening this week so if you happen to be in Amsterdam, reach out to us.
Join the conversation on social media: Follow our twitter account or use the hashtag #WSAI17 ❤️

Applied #AI, by Azeem Azhar
China doesn’t boast about its dentists. The country is short on trained dental professionals and surgical errors are frequent. To fill the gap, Beihang University in Beijing and Fourth Military Medical University developed a robot to perform first fully-automated dental implant, programming the machine with tolerances of less than 0.3mm and movement detection.
This is the latest impressive example of technology moving into the workplace of once human-only profession. Even with its limitations, some AI systems can predict heart attacks more accurately than human doctors. It performs legal research faster than any human. Algorithmic lip readers outperform us, deciphering twice as much information. Digital interviewing systems are screening applicants for jobs in investment banking, considering thousands of different traits—an unimaginable feat by the human counterparts. The list goes on.
A recent survey among more than 300 machine learning researchers revealed that it’s likely AI will outperform humans in all tasks in 45 years - with Asian researchers expecting it sooner than North Americans.
Researchers predict AI will outperform humans in many activities in the next ten years, such as translating languages (by 2024), writing high-school essays (by 2026), driving a truck (by 2027), working in retail (by 2031), writing a bestselling book (by 2049), and working as a surgeon (by 2053).
Humans are flawed in prediction, and the lack of consensus in expert opinion captures this. Our robo-dentist successfully mastered one procedure, but the fact of the matter is machines are still not fit to deal with unpredictable work, or work requiring adult perception and mobility.
In 2010, researchers at Berkeley University revealed a towel-folding robot which it took an average 20 minutes to fold one small towel. Some progress has been made: there’s a startup offering an oversize laundry-folding robot that de-wrinkles and folds clothes in about 40 seconds. However, users have to forget about folding socks, underwear, large towels or XXL clothes.
Adult brains, but underdeveloped sensorimotor skills are the signs of Moravec’s paradox. The paradox is echoed in a recent McKinsey study, as well: 78% of predictable work will be automatable. It turns out the largest proportion of these types of tasks are in manufacturing. Data collection and processing are close followers: impacting roles in financial services and retail. By contrast, applying expertise, managing people, working in unpredictable environments, are on the lower end (safer if you will), made up of 9-25% of automatable tasks. These tasks require complex communication skills, empathy, and ideation — at which most humans fare quite well.
As the leader in warehouse automation, Amazon offers clues about the (near) future of workers. So far, the company claims that no worker has been laid off because of automation (within Amazon), while employees report their new jobs are more “mentally stimulating”. Some have been retrained to babysit the machines, others to do delicate jobs that robots are too clumsy to perform. In fact, Amazon is in warehouse hiring spree, expecting to fill in 100,000 new roles in the U.S. On the other side, many hundreds of thousands of retail workers will go down with the rise of e-commerce in the coming years, through the courtesy of Amazon.
A precursor to Amazon’s lesson in re-developing jobs within the company was the introduction of teller machines. ATMs didn’t eliminate bank clerks. Clerks settled into new jobs within the banks.
Since 1980’s, large fraction of US employment growth has come from new jobs, according to MIT’s Acemoglu and Restrepo. In the 80’s management analyst was a new title, in 1990’s it was radiology technician; in 2014, cloud developer became a thing. In the coming decades, new categories of jobs are emerging. We’ll soon see a need to fill in for what Accenture’s Paul Daugherty calls AI trainers, sustainers, and explainers as the work of the machines powers new opportunities.
It is misleading, however, and philosophically falsifiable, if you will, to claim that we know what’s going to happen in the long run because we’ve seen it all in previous technological revolutions. We can’t predict the future accurately relying on the past.
In the coming wave of technological unemployment, it seems intelligent systems will replace not only blue collar workers but also white collar, previously with high wages and higher skills. Collaborative robots that work alongside humans are a thing now, but are they still going to need human babysitters in two or three decades?
As machine vision improves, and combines with more balletic sensorimotor performances, where will hundreds of thousands of Amazon employees, now in their 40s and 50s, be redirected to next?
One possibility may be their humanity. However imperfect their accuracy, vision, speed and versatility relative to a system touting machine intelligence, their provenance as human will be perfect. (No machine can ever have a perfectly human provenance. Not at least, until humans become machines.)
In that future world, robot-made will be ubiquitous. That high supply of goods and services borne through automated, mass-available, 24/7, highly efficient systems will drive their price down. Somewhere, still, human imperfection will gain value. One trait humans all the way longed to hide and exterminate, will now be the savior of human purpose. Food, art, clothes, furniture and else will be a luxury if touched by human hand — the hand so prone to making mistakes. I can imagine going into a supermarket and seeing on the shelf products labeled “not touched at all by a robot or machine.”
This is artisan economy. Humans will lose their repetitive warehouse jobs, but we’ll get an opportunity to excel at intimate and experiential. At least, some will.
Many of us might not adapt. Taking crumbs of jobs with lower dignity, pay, and security. Or the economy might only support so much artisanal cheese or hand knitted sweaters. If that’s the case, a new social consensus for allocating society’s rewards will be required (whether it is Universal Basic Income or something else). Complete unemployment draws, perhaps more complex, uncertainties other than income. Work is a socialization process, part of one’s identity. How are we going to fill this gap? There are plenty of questions like this that we have no answer to. But, there are even more questions that we yet don’t know we need to ask.
TechCrunch Disrupt Berlin 2017
AI Shines At TechCrunch Disrupt Berlin
TechCrunch Disrupt is coming to Berlin on 4-5 of December. Come and discover early-stage AI startups and hear from leading thought leaders in the field like Stan Boland the CEO of FiveAI and ABB Group’s Ulrich Spiesshofer. You’ll also be able to attend the Artificial Intelligence “Off The Record Session” where you can interact with and ask questions of AI-focused speakers and disrupters. Book your tickets today and save 15% with promo code “WORLDAI”. 
See you at World Summit AI!
Special thanks to Azeem again! Follow his excellent weekly newsletter on the link below:
Would you like to contribute to our newsletter, or have something interesting to share? Tweet us @thecityai or shoot us an email!
Sharing is caring! Share via Twitter 🔥 | Follow us on Medium
Curated with   ❤️  from London,
Azeem & The City AI team
Did you enjoy this issue?
City AI
By City AI

Applied Artificial Intelligence - Local AI news - Events - Use Cases - Global Collaboration between 40+ cities

In order to unsubscribe, click here.
If you were forwarded this newsletter and you like it, you can subscribe here.
Powered by Revue