1. Artificial intelligence is coming your way. To protect your career, it's best to take jobs that machines aren't good at—jobs involving people, unpredictability, or creativity. Avoid repetitive or structured jobs that will automatically disappear in predictable environments, such as telemarketers, warehouse workers, cashiers, train operators, bakers, or assembly line cooks. Truck, bus, taxi, and Uber/Lyft drivers may soon be replaced. Beyond these, many other professions (including paralegals, credit analysts, loan officers, bookkeepers, and tax accountants) are not on the list of those to be replaced, but much of their work is automated and therefore requires fewer people.
2. Unemployment could become a lifelong vacation. Advances in artificial intelligence could either create a luxurious and leisurely society for everyone, or bring unprecedented suffering to the majority who are unable to work, depending on how the wealth produced artificially is taxed and shared.
3. Killer robots are not fictional. We are on the verge of an out-of-control arms race in artificial weapons, which could weaken the military power of today's great powers because it would allow everyone, including terrorist organizations, to possess a cheap and convenient assassination machine with a full wallet. Artificial intelligence researchers oppose this and hope for an international treaty on the control of artificial intelligence weapons.
4. Machines do not possess intelligence. Intelligence, on the other hand, is necessary for accomplishing complex goals. It cannot be quantified by a single number, such as IQ, because different organisms and machines excel at different things.
5. Artificial intelligence is becoming increasingly widespread. Today's AI primarily refers to intelligence in a narrow sense—the ability to complete a set of small goals, such as playing chess or driving, sometimes performing better than humans. In contrast, humans possess true intelligence—the ability to accomplish any goal, including learning. The ultimate goal of AI development is Artificial General Intelligence (AGI), the ability to perform any intellectual task like humans. Many leading AI researchers believe we are only a few decades away from AGI.
6. Artificial intelligence may leave us far behind. As British mathematician Irving J. Good explained in 1965, "A superintelligent machine is defined as one that can far surpass human intellectual activity, however intelligent it may be." Since machine design is an intellectual activity, a superintelligent machine can design even better machines. Undoubtedly, this will be an "intellectual explosion," and human wisdom will lag far behind. Therefore, the first superintelligent machine is the last invention humanity needs to create, provided that the machine is docile enough to tell us how to control it.
7. We are far from reaching the limits of computation. The cost of computers has decreased dramatically since my grandmother was born. If everything became cheaper, one percent of that money could buy you all the goods and services produced on Earth this year. Moore's Law dictates the cost of moving electrons on a two-dimensional silicon wafer, and once we get to that platform, we can try many other hardware solutions—such as using three-dimensional circuits, using electrons to complete our bidding. We are still a quadrillion times lower than the limits of computation according to the laws of physics.
8. Artificial intelligence can help humanity prosper. Because everything in human civilization is a product of intelligence, by amplifying our own wisdom with artificial intelligence, we can potentially help life flourish as before and solve our most intractable problems, from disease to climate change.
9. Artificial intelligence brings risks. Hollywood's fear of machines becoming conscious and evil is a distraction. The real concern isn't malice, but capability. Superintelligence is defined as being exceptionally good at achieving its goals, whatever they may be, so we need to ensure its goals align with ours. Humans generally don't dislike ants, but we are far more intelligent than them—so if we want to build a hydroelectric power station, and there are anthills there, it's simply terrible for the ants.
10. We need AI security research. To ensure AI's growing impact on society, more research into AI security is needed. For example, how can we transform today's vulnerable computers into truly trustworthy and powerful AI systems? How can we enable machine learning, acceptance, and retention of our objectives?
These are challenging questions that may take decades to answer, so we should begin research now to ensure we have the answers when needed. But compared to spending billions of dollars to make AI more powerful, governments have so far allocated very little funding for AI security research.
For more information, please visit the Industrial Robots channel.
