Stephen Hawking recently published a brief op-ed on artificial intelligence (AI) in the Independent. In it he discusses possible implications of rapidly developing AI and the need to prepare for the changes it will bring. Media decided to put its own twist on the matter. In the past week we have been inundated with headlines like, “Stephen Hawking Is Terrified of Artificial Intelligence”, and “Stephen Hawking: Artificial Intelligence ‘Potentially the Worst Thing to Happen to Humanity’”. Is the great Stephen Hawking really the neo-luddite that these headlines seem to suggest that he is? Is he really warning us about a technological leap that should be avoided at all cost?
No, nothing could be further from the truth. I think Hawking makes it clear in his article that what is to be feared is not AI or other technological developments, but rather our (as in ‘we humans’) lack of foresight and planning to ensure that technology develops in a way that produces the greatest benefits for humanity. Although this applies generally, one of my big worries is that education is especially lagging in this regard. Not only are we woefully unprepared for the impacts that AI and other technological developments will have on the field of education, but I think that relatively few are even aware that the potential for significant impact is even there, and closer than many think.
So, here are what I think are several overlooked aspects of AI and related technological developments that educators should be preparing for:
1. Self-driving cars – Not many educators are likely to see any immediate connections between the development of self-driving cars and education, and for the most part, they’re right. However, it is clear that the powers that be have envisioned a near-term future (next 5-10 years) where self-driving cars will become the norm (see pg. 7 in link) and are vigorously pursuing ways to make this a reality. This means, among other things, that advanced AI technology needs to happen and it needs to happen quickly. However, advanced AI developed for self-driving cars will not only serve the auto industry. The technological hurdles that need to be overcome on the path toward viable self-driving cars, including advanced AI, will fuel developments in other areas and will do so very quickly. Whether educators want it or not, advanced AI will be in, and affecting, the learning environments that they construct around the time that children being born today start school.
2. Superdupercomputing – Effective AI requires a lot of computing power, in fact massive amounts of it, the kind of power that only the most super of supercomputers can supply. In our minds, supercomputers are something that you find in advanced science instutes; not something that we, the lowly commoners, interact with. That is wrong. Much of what we do with information and communication technologies today involves some of the most powerful supercomputers being deployed today. We are no longer beholden to the limitations of the device we have before us. The increasing availability and accessibility of high-speed data networks allows us to pass on many of our intensive processing tasks to supercomputers. In the past, when we heard about the vast processing power of one or another government-funded supercomputer, it had little relevance for the everyday technology-user. Today, when we hear about Watson’s or other supercomputers’ amazing new capabilities, it means something is about to happen that will affect us.
3. Computer-generated information – Computers are data-processing machines. They are very good at sifting through abstract data, identifying patterns, sorting, and that type of stuff. Making sense of patterns in abstract data and being able to communicate that in a meaningful manner is a whole other ballgame. Humans are far better at making sense of stuff than computers that see abstract data as meaningless tokens. Increasingly advanced AI is challenging our predominance in this area. Already, media outlets are using various software solutions to generate brief, concise news items from abstract data far more quickly than humans can. There are already a range of services and applications that generate anything from brief news items and blog articles to extensive reports and books. A recent study showed that while computer-generated articles tended to be somewhat dry, participants in the study were not able to distinguish between them and articles written by humans. If you are a teacher, it may very well be that you have already received from a student some sort of paper that was entirely generated by a computer. The accessibility of these types of services will increase as AI technology develops.
4. Change faster than the speed of the human mind – As computers increasingly take on the task of generating information from data, rates of change in various knowledge domains will increase to levels that are beyond our ability to keep up. This is already happening in some areas. The increasing use of software to facilitate financial transactions is perhaps one of the best current examples. It is estimated that computer-driven high frequency trading (HFT) already acounts for around half, or more, of all US equity trading volume. With modern computers and complex trading algorithms, trades of this type occur in microseconds. Consequently, significant market fluctuations can occur before we mere humans would even be able to figure out what happened, nevermind formulating a reasonable response. Use of automated article-writing software is creating similar circumstances for journalists who thrive on “breaking the story”. Computers are able to grab data, put together a concise news report and publish in the time that it would take us to get to our computer and orient our fingers on the keyboard. Just a month ago, the Los Angeles Times broke the story about a major earthquake in Southern California just minutes after it occured. The news item was generated by reporter and programmer, Ken Schwencke’s, Quakebot, software that he created that monitors U.S. Geological Survey’s communications and writes up a story the moment a significant event occurs and automatically posts it to the LA Times LA Now blog.
These all pose significant challenges for educators and others involved in education that need to be addressed. But, each also offers possibilities of fantastic opportunities, as long as we prepare ourselves. The threat here is not in the technology, it is a question of how prepared we are for some pretty dramatic changes.