On March 1st Microsoft president Brad Smith gave a guest lecture at Princeton University about the role of Artificial Intelligence (AI) in our society. Referencing Microsoft’s recently released book, “The Future Computed”, Smith focused on how AI will impact jobs and the economy.
“It is really not helpful to think about technological change in terms of what is going to happen in a year or two years…It really helps to step back in terms of ten or twenty years,” Smith said.
He asked the audience, comprised mostly of Princeton students, to look back on the world twenty years ago, and think of the technology that was then in use—noting that digital technology was in little use, unlike it is today. Since then technologies such as the smartphone and the laptop have become part of our daily routines, suggesting that other new technologies will emerge over the next twenty years.
“I think it is fair to assume that it will probably be grounded in artificial intelligence,” Smith said. “We are already living with so-called digital assistants whether it is Siri or Alexa.”
Digital assistants will grow and become a part of our daily lives, according to Smith. Many of us will eventually engage with a digital assistant when we start our daily routines, rather than a radio.
“The real goal in our view is to make sure that all of this technology actually helps us and amplifies human ingenuity,” Smith said.
Though artificial intelligence is already part of our lives in the form of the consumer products and services Smith referenced, he stressed that it is more important we view AI in terms of jobs and the economy, which requires a deeper understanding of what AI is.
He defined Artificial intelligence as “the ability of machines to take in data to learn and to make decisions. But it is also many things, and one way to think about it is that it is the sum of the many pieces that go into AI.”
For instance, AI occurs because machines have the ability to see, as a camera sees an image. But the difference now is that machines have the capability to discern what the image is, Smith explained. The same is true for speech that machines can understand what is being said, but advances have now made it possible for machines to understand the meaning of what is said. Ongoing improvements in these individual fields will lead to further advances in AI.
“We expect there will suddenly be one day when it (AI) is all here… all at once. But that is not the way it works. Each piece advances and we are already living in a world where AI is around us…in pieces,” said Smith. He gave the example of the technology in use by BMW that enables the car to sense what is happening and then make the decision to apply the brakes—without first engaging the driver.
This led Smith to ask a key concern about AI, “When does the computer get as good as an average human being?” He then offered real life examples of computers taking over tasks currently done by humans.
Meetings now being held where the participants speak many different languages require translators, but in the future, computers will translate the presentations in real time for the many and varied languages present in the room. Human translators will no longer be needed and this may be reflective of what will occur in other industries.
“All of these things are made possible by two fundamental advances in computing. They are the fundamental engines of innovation for every single thing,” Smith said, describing the key roles of huge computational power now available through large scale data centers and the availability of large amounts of data now being insatiably collected. These two areas will drive AI forward for the next decade or more.
“As computers behave more like human beings, what is it going to mean for us, for real people?” Smith posed. He feels this is the time to have a critical eye about the changes being brought about by the technological advances in AI. He offered two ways to approach the future.
First, one should learn from the past, not using it as a predictor, but rather discerning trends as a way to gain insight to be applied in the future. To support this Smith gave as an example the advent of the automobile industry that began in 1885 when Karl Benz invented the world's first 'production automobile’ that used an internal combustion engine.
Over the next 20-30 years the automobile industry displaced the existing horse carriage industry. But it also spurred other innovations such as the invention of consumer credit to support the purchase of higher priced automobiles and the development of corporate logos to be used in advertising for a nation on the move and passing by billboards where logos could stand out.
But, the adoption of the automobile triggered economic changes that played out over decades, according to Smith. The disappearance of carriage manufacturers and the decline of horses triggered farmers to change their crop production that in turn dampened agriculture prices and depressed farm incomes. This in turn led to wide-scale foreclosures of farms by world banks that led to a broad financial collapse. So an innovation that began in 1885 disrupted industries causing whole scale unemployment that rippled around the world and decades later had a role in bringing about the Great Depression.
While other innovations emerged that were helpful during this period, Smith believes the real lesson is to be prepared for the unexpected and to be agile when forces of such magnitude appear. That is the approach to be applied to artificial intelligence that has begun to impact our society.
Smith’s second lesson is to believe that more opportunities will arise associated with these new innovations and industries--- because historically that is what occurred in previous technological disruptions. As jobs and professions in industries impacted by technological change disappeared, other industries emerged that created new jobs.
Taking an analytical approach as to what computers do well, he identified the core areas of speech, vision, language, and knowledge. He then applied that to identify the professions that will likely be adversely impacted such as radiologists, call center personnel, fast food order takers, auto and truck drivers, translators and interpreters, machinery inspectors and repair people, and paralegals.
But, he then offered a profile of the types of jobs that machines are not good for such as those where human creativity, human empathy or human interaction is important. This includes nurses and social workers who will use AI, rather than be replaced by AI.
“We need to ask what jobs AI will create,” Smith stated, noting that it is easier to identify the jobs that will be lost. Predicting what new jobs will be created is a challenge that is magnified when one hears of the studies about AI and jobs. They project that those entering kindergarten today will face an economy where 65% of the jobs currently in existence will be gone when they graduate from high school..
Though that seems to be daunting, Smith feels that there are other pressing issues beyond where the jobs will be created---and that applies to the companies creating AI, moving Smith to discuss the role of ethics and AI.
“It is not what computers can do, it is what computers should do,” Smith said. He feels that this is one of the great questions of our time and it was part of Microsoft’s thinking when they put their book together about AI.