Artificial Intelligence in the Near FutureAdd bookmark
Have you seen the movie Short Circuit. It's a fantastic 80s movie about a robot who gains individuality and intelligence after being struck by a bolt of lightning. The robot, named Johnny Five, is immediately quotable using the catch phrase, "Need input."
While the movie was certainly meant to be a comedy/scifi film, the reality is artificial intelligence works the same way. It needs input to function. AI also needs computing power. In fact there's been such an increase in computing power that suddenly movies like Short Circuit aren't that farfetched. In the last five years alone, the boom in computing power has allowed algorithms developed around 20 years ago to actually be used.
"The neural networks and algorithms that run Watson and other AI applications actually work now because the computing power is there to allow them to," explained Robert Clapperton. He's over product development with Ametros Learning.
But it's not just about ones and zeroes; algorithms and computer science. Combine the input and the computing power and suddenly you have a machine capable of learning. Companies like Google, IBM, and Microsoft are already heading down that road.
"The work they're doing… once you start the programming and you create the algorithm, the algorithm starts to the function and the more data you feed it, the better the algorithm becomes and the better the machine learning becomes," Clapperton said. Clapperton explained that translates into the AI machine learning at a very rapid rate.
As AI machines continue to learn, so do humans about its applications. AI has taken the HR space by storm. In fact, the technology is already being used to narrow down applications based on a pre-programmed set of criteria. Some are using it to power chat bots. Chat bots allow potential and current employees to ask questions and get a response based on the question. As an example, let's say an employee has a particular question focused benefits. A chat bot is able to search the query and answer it for the employee without any assistance from an HR professional.
While that seems useful, the reality is Artificial Intelligence is "guessing" at the response. What does that statement mean? AI is able to search a query based on the words you are using and give you a response, but that response isn't contextual. Clapperton says that's where AI is heading now.
"Instead of writing responses specifically to specific inputs… you just have a huge database of language around a specific knowledge domain and the AI can go into that knowledge domain and answer the questions of the user," Clapperton said.
Educating via AI
HR professionals interested in pursuing AI want it to do much more than answer questions and rummage through applications. They want to use it as a learning platform.
Clapperton says it's not there yet. AI can teach itself to do something, but it's not at the stage it can replace humans beings as the "drivers of education." He points out in the future it may be used that way, but it would require a lot of adaptability.
"I think the adaptability of content and what's being taught is going to keep getting better as more and more data sets become available so that you can have the data to do the type of machine learning on to generate responses that are specific to what a person actually needs to learn," Clapperton said.
He gave the example of an AI-controlled multiple-choice test. If a person taking the test marked an incorrect answer, AI would then give them a question a bit easier to answer. If the question was answered wrong again, AI would follow with a question lower in difficulty level. When the student began to answer questions correctly, the difficulty of the questions would increase. Similarly, a person answering questions correctly would continue to get more difficult questions. This allows the AI to determine what topics the student understands least. In doing so, learning becomes personalized and specific for the student.
Of course, with any learning tool trainers and students must be open to it.
"I think the people that understand that it is a tool to help train people more effectively, and especially more scalable in a scalable way like… that's the biggest thing AI can do. Then you can create these simulations that can be done by thousands of people," Clapperton said.
And that's just what AI can do on its own. Imagine if it's paired with another technology such as virtual reality.
Clapperton says AI has a big role to play in terms of VR training. AI can be used to make better training experiences by collecting data about each experience. That's just the tip of the iceberg. It can also be used to add a communicative piece to a VR training simulation.
"You may be learning how to use a crane, but at the same time, you're talking to foreman and you're talking to other workers and you're communicating with the people on the ground as to how they want to move this material around and everything else. You're adding this whole other layer of simulation to the training, so it's not simply a matter of can you work the controls to lift something up and move it around… its are you able to communicate effectively," Clapperton explained. "The beautiful thing about AI training like that is you can you can program the AI to create challenges that they may face in the world, you know, when they're actually operating the crane. AI definitely adds that communicative, interactive layer to it and allows you to train the kind of intelligences that are really hard to train… like emotional, and social cognitive analytical. That sort of thing."
Robert Clapperton is with the Product Development Department at Ametros Learning.