The Sound of Science - 'The Future of A.I.'

Jan 3, 2020

Sam: Welcome to the Sound of Science on WNIJ, I’m Sam from NIU STEM Outreach

Fred: And I’m Fred Williams from NIU Division of Information Technology. I’m not an expert in Artificial Intelligence, and I don’t think anyone can truly predict the speed and breadth of AI development. However, I have some ideas as to where AI is going.

NIU STEM Outreach

Sam: Let’s start with how AI is already guiding our lives. AI is really good at finding patterns and making predictions given enough data. This is called machine learning, and it’s where a program is fed a bunch of starting data and the results. This makes AI very adept at spotting anomalies. For example, a bank already has your name and address as well as the data on what kinds of things you buy, where, when, and how often. If you buy groceries in Chicago then all of a sudden your card is used to buy a camera in Beijing, the bank is going see that as an oddity and shut it down.

Fred: AI can also help defend against cyber-attacks. If someone’s login credentials have been compromised and an attacker is snooping around, then an AI program might say to itself “This user doesn’t venture into the server at this time of night and certainly not into these systems and files. Let’s lock this down!” The AI can then perform predetermined tasks like completely blocking out the fraudulent user, alerting the security team, or locking down servers.

Sam: But as AI develops, there are ways it can be used to speed along time consuming tasks. For example, one proposed use is to feed an AI program all the models and platforms of a video game in development. A developer can then tell an AI “Using these premade models, make me a landscape this size with that many features” instead of doing it all by hand. 

Fred: Or a computer programmer can order a handful of skeleton codes to use as building blocks. In all of these cases, the AI is not thinking on its own, it’s still following a specific set of commands based on a programmer’s inputs. Nevertheless, it’s creating building blocks and foundations at a much faster rate and in predictable ways.

Sam: Again, these are just some possibilities. There is plenty of room for growth and expansion, such as in the realms of self-driving cars, food production, healthcare records, and marketing analysis. 

Fred: Hey, you spit those ideas out pretty fast… that makes me wonder…

Sam: Thanks for your time Fred! This is the Sound of Science on WNIJ

Fred: Where you learn something new everyday.