I am glad that the festive time of year has once again drawn to a close. The bounty of turkey and mugs of Irish coffee got to be a bit much for me. The time with friends and family did, however, gift me some interesting ideas.
One idea came from my uncle who holds a senior position in IBM. He told me about their machine called Watson, described as “a cognitive technology that can think like a human.” It can understand and, most importantly, learn from human emotion and language.
Watson’s abilities got me thinking—what is the current state of Artificial Intelligence (AI)?
The field of AI research was founded at Dartmouth College in 1956 by a group of cognitive scientists. Plenty of advances in technology have happened since, but most progress in AI was made in the last decade.
Today, Microsoft, IBM, Google, and Amazon are pouring billions of dollars into AI. We are now seeing the payoff in big advances in frontiers like machine learning.
Machines like those in factories are programmed by the user to perform a unique task, nothing else. With learning machines, there’s no need for exact instructions. Instead, the user defines an end goal, and the machine learns how to achieve the goal through trial and error.
Machine learning is behind advances like driverless cars, image recognition, and unbeatable Go players. With some of the world’s most successful companies now focused on AI, how might its progress affect us?
Humans Need Not Apply
A 2016 report by McKinsey looked at the ability of machines to replace human labor. The report found that 59% of all manufacturing tasks could be automated using current technology. The most exposed sector is food service/accommodation, 73% of which could be replaced. The least exposed was “managing others,” where only 9% of tasks could be automated.
The financial sector also has cause for concern. McKinsey found that financial workers spent 50% of their time collecting and processing data—easily automated. For proof, just look at the floor of the NYSE to see how machines can control a space once alive with activity.
Machines have slowly replaced factory workers over the last century. In today’s new era, AI could rapidly replace human labor across several industries. A 2013 research paper from the University of Oxford found that 47% of US jobs could be automated over the next 20 years.
Ford wants to have driverless trucks on the road by 2021, and Uber already has autonomous taxis. Drones also look set to affect the transport industry and replace security jobs. And prepare for lattes served by robot baristas in coffee houses across America.
Many of the most vulnerable jobs are filled by low-income and middle-class workers. The changes will hit them hard.
This would create major headaches for policymakers. How could millions of people be quickly retrained with new skills? The retraining programs themselves could become obsolete given the rapid rate of change. Skills being taught in universities today may be useless within a few years.
While the rise of AI will certainly create challenges, it will open up a captivating future.
Automation will be both disruptive and liberating. It will free up humans to do more productive, advanced tasks that could lead to great prosperity. Think America in the early 20th century, on steroids.
Along with productivity gains, AI has the potential to revolutionize sectors like education and healthcare.
AI systems have just begun to grasp human language. In time, this could make personalized education feasible. For a struggling student, AI could slow down the lessons or change teaching style. If the student excels, AI could raise the difficulty. All things being equal, this would enhance human learning.
In healthcare, IBM’s Watson is being used in an oncology program at a cancer center in New York. Watson was tasked with diagnosing 1,000 oncology cases that had been handled by field experts to see its potential. The results were promising. In 99% of cases, Watson drew the same conclusions as the experts. In 30% of cases, it found something new that the experts missed.
Watson recently helped a woman with leukemia who had been misdiagnosed. By analyzing her genomic sequence, Watson was able to tell doctors that they had missed a second strain of the disease. Treatment was adjusted, and the patient recovered.
Google and Microsoft have both recently made discoveries that could change the field of speech technology. IBM is working on bringing AI to the blockchain—a core component of Bitcoin. In 2016, Barclays carried out the first trade transaction using the technology.
While progress thus far is encouraging, we have only started to see what is possible. With Peter Diamandis’ and IBM’s XPRIZE partnership and Elon Musk’s non-profit OpenAI, resources will flow into AI.
With as many hopes as fears, what can we take away from AI’s rise?
Splitting atoms can create huge amounts of energy, but also destructive weapons. AI shares this same good/bad paradigm. There is no doubt that AI can be highly disruptive to social and economic structures, and potentially lethal. It also has the potential to transform the world as we know it.
Anxiety about vast replacement of human labor across industries is a serious issue. Millions of workers rendered obsolete over a short time span is a recipe for mass social unrest. No doubt there will be great need for upskilling in the years ahead.
There are further concerns about AI becoming so smart that it goes rogue and carries out destructive acts. Andrew Ng, a leading computer scientist, says such fears are like worrying about the overpopulation of Mars. AI advances so far have been narrow AI. The development of general AI, a super-machine if you will, is a while away. We don’t have to worry about Hal 9000 anytime soon.
Some of the world’s greatest minds are working on AI, and we should be optimistic about its future. AI could be the biggest game-changer in history. Its development brings possibilities that we cannot yet imagine.
The potential for AI to better humanity seems infinite. For the human vs. robot debate, I think partnership, not replacement, is the likely outcome. After all, who would bet against human ingenuity?
Once again, thank you for tuning into this week’s The New Abnormal. If you have any questions, or topic suggestions you can connect with me on Twitter.
And with that, I’ll sign off for this week.
Chief Analyst, RiskHedge