A student at the Thomas Jefferson High School for Science and Technology in Alexandria, Virginia, hosted a nationwide summit on the subject of artificial intelligence on September 30, writes Sneha Joisha in tjTODAY, the student newspaper at the high school.
“I feel like giving students a foundation, especially in artificial intelligence, is really important,” the paper quoted Kavya Kopparapu as saying. The senior was a semifinalist in the Siemens competition. “I’ve attended a lot of research conferences presenting my research, and [for] a lot of them, I’ve been the only high school undergrad or graduate student because it’s mainly for professionals. And so bringing the atmosphere of that type of conference and all the amazing speakers who are industry leaders and bringing them so that high school students can interact with them [is important].”
To the summit, which took place at a local hotel but was available to students from all over, she invited several big-name keynote speakers, including Johns Hopkins University Assistant Professor of Finance Jim Liew and members of the Girls Computing League, which she herself founded.
“I’ve lived such a life where computer science has been taught to me since I was in preschool, but it just occurred to me that other kids don’t even have access [to] the camps that we’ve done,” a junior who attended the summit was quoted as saying. It “just struck me as how much we take for granted [and] how much we have the Wi-Fi [to] code, and others don’t do that.”
While each of Ms Kopparapu’s guests had a few words to say about AI, Voxitatis also has a few to add. Whether we want to admit it or not, computers are being used for scoring essay responses on standardized tests in just about every state. The reason is that it’s less expensive for states to pay a computer to give scores to thousands of responses an hour than it is to pay humans to do about 50. Plus computers don’t require health insurance or other employee benefits. They don’t complain or sue the company. They don’t harass other workers. And well, you get the idea.
Data is an essential backbone of any organization in existence today, including schools and state departments of education that rely on data from standardized tests. The thing is, there’s so much data that humans can’t get a handle on it quickly enough to make any useful decisions based on it.
Let’s take another example. A recent report from General Electric said that more than 10 terabytes of data is generated by each engine alone for every 30 minutes a plane is flying. A human just wouldn’t be able to read that much data, analyze it, and use it to guide decision-making about the flight. But a computer can. It has no problem with speedy decision-making if it’s well programmed.
This is what so many people fail to understand. AI is not a computer program; it’s an “intelligence.” There isn’t some big program or flowchart out there that says, If you get this value, do that or the other thing. Computer programs work like that, like a simple machine.
Put enough coins in a vending machine and press the button; a can of soda pops out. That’s because putting in coins changed the state of the machine, and if the state of the machine is such that a sufficiently large amount of money has been inserted, it’ll give you your Coke.
AI is a very different animal, as it doesn’t have some preconceived notion or pre-programmed flowchart telling it how much money should make the vending machine deliver its product. Using AI for such a problem would be completely inappropriate—it would take too long for the intelligence to learn how much money was enough, because it would have to be taught how other vending machines behaved, analyze the economics of the whole price-setting strategy, etc.
AI computers get smarter as actions are repeated. They learn from their mistakes and from their successes. They can be trained based on the success of other machines or of humans who perform similar tasks. Like Siri, Cortana, or Google Assistant, AI computers can process spoken speech and, via natural learning processing, analyze that speech. Even if you use slightly different words, the computer can figure out what you’re talking about and predict a logical course of action for you to follow.
AI might replace jobs
Just as computers can replace human scorers on standardized tests, AI has the potential to replace about 7.1 million jobs by 2020, the World Government Summit found. Image-processing and voice-recognition computers are already making inroads in certain fields, such as routine manufacturing, but even if a computer doesn’t replace the human, it can make the job that human does better—as with the airline flight. Or how about all the image data the human mind has to comprehend and act on during a heart surgery or a brain surgery?
Programming AI computers will also create new jobs, and some of these jobs look very different from the ones people have today. They involve not only statistics, computer programming, design, and engineering, but also the soft skills of working well on a team, interpersonal communication, and so on.
These “soft” skills, which also include logical reasoning, common sense, and creativity, are not easy for a computer. In fact, most people in the AI field consider these skills outside the scope of what computers can even attempt. That’s where the jobs are going to be for humans for a long time to come, I would think.