Opinions expressed by Hawking, Gates, and Musk about the dangers of artificial intelligence rang loud and clear in 2015, and continue to echo into the new year. Since then, there have been plenty of predictions of humanity’s doom at the hands of autonomous machines. But there have been leading thinkers in the AI space who have also come out from behind the curtain to play devil’s advocate and make clear opposing positions.
In a recent TechEmergence interview, Author and Artificial Intelligence Researcher Dr. Steve Omohundro says that we’re still in the very early stages of AI development and future directions still lie in the decisions that we make today.
According to Omohundro, technology is driving a major transformation in society. To illustrate, he cited estimates that $50 trillion of value will be created by artificial intelligence and robotics in the next 10 years, and that one third of all jobs will be automated in the same period. One industry that is on the verge of big changes thanks to AI, he said, is the legal profession.
“I was just at a future law conference in which they were all talking about the legal profession being completely transformed by AI. Every aspect is going to be transformed by the technologies,” Omohundro said. “There’s a transformation in the legal industry where much of what human lawyers do today is really quite routine… like creating a contract, yet it’s very highly paid. So there’s a tremendous incentive to automate as much as possible of that aspect of the law.”
Another area where technology is changing the legal world is Ediscovery, in which law firms are applying natural language technology to sift through what could be millions of a company’s emails during the discovery process. Beyond that, new technologies such as self-driving cars are creating a variety of new ethical questions (MIT Technology Review’s Will Knight provides a glimpse into the philosophical considerations of this issue) and challenges on which the legal system has yet to gain a handle on or even a full set of potential ethical situations.
That’s just the tip of the iceberg, he said, as the largest impact of artificial intelligence and its legal ramifications will impact society at large. “The area that, at this moment, is having the biggest effect is big data and machine learning being applied. All the consumer companies are gathering all kinds of data about their customers, predicting what kinds of ads are going to (sic) be most effective and targeting that kind of information,” Omohundro said. “How do you manage that? What are the legal restrictions on that?”
As artificial intelligence continues to develop in the future, Omohundro sees three areas that will be affected the most - the economy, the military, and the potential for super-intelligent-based singularity as theorized by Nick Bostrom. We’re already seeing AI’s impact on the economy, he noted, as businesses are looking at things that are inefficient or can be done better by automation, and realizing significant profit can be made by eliminating those inefficiencies, he said.
In the military sector, while drones are already a well-known innovation, Omohundro notes that in the United States, there’s always a human making the “kill decision.” Yet that may not be true in other countries, and the potential of a more divergent country developing its own autonomous robot army should not be taken lightly.
“The third wave is when these systems become of the same kind of intelligence or more intelligent than humans. What does that do to society? If in 10 years, 30 percent of all jobs are all automated, what does that say in 50 years?” Omohundro said. “Are humans going to (sic) be able to work at all? What is the role of humanity in that kind of future? I think there are huge, huge questions there that we’re just now beginning to address.”
Omohundro compared the potential for super-intelligence to the nuclear arms race. While we went through a period of time where such weapons could have easily been used toward the destruction of the human race, most countries have realized their danger and implemented agreements to curb their use and development. The same thing will need to happen to ward off the danger presented by autonomous-level AI, he added.
“I think this is probably gonna’ be the biggest challenge that humanity has ever faced. The realization that ‘Oh, this could be a very powerful technology,” I think that’s just dawning on us now,” Omohundro said. “We’re building these things, there’s nothing inevitable about any of it. We should think it through very carefully and design systems that actually reflect the kind of future that we want.”