As a leader in digital innovation, Kansas City regularly explores the idea behind AI. 1 Billion Bits came together to discuss implications of the infrastructure necessary for Artificial Intelligence during their quarterly lecture / panel series on Wednesday, September 27.
Artificial intelligence is the simulation of human intelligence by machines, which includes learning, reasoning and self-correcting. Programmable automation, machine learning, deep learning, machine vision and natural language processing are all encompassed by the term Artificial Intelligence. This means that things such as cars, financial analysts and manufacturing call all be automated with a self-learning system such as AI.
The infrastructure necessary is still up for interpretation, the panel decided. The only constants with artificial intelligence are the necessity for low latency networks, increased data storage and strong security. This is nothing new — of course robotics that absorb, store and analyze tetrabytes of data need more of everything:more speed, more storage, more security. But it’s the methods that vendors will deliver to the cities housing this new infrastructure that is the question.
“We have a basic network, it’s the internet. Now we are overlaying that network to get you somewhere specific,” Jon Wagner said.
Things like latency can be solved a number of ways. Take self-driving cars, for example. Shortwave communication is shaping up to be the most economical and safe way to drive with AI, and thus infrastructure would need to be built and maintained for short radio waves among most streets. While roadway infrastructure is tedious as it is to build, storage solutions for the massive amounts of AI data seems more insurmountable. Though big data repositories have started to appear, there are still questions of payment and data use.
“I can collect all this data, but where do I put it?” Nate Olson-Daniel of Light Edge Solutions said Wednesday night. “And what am I using it for?”
What the data’s being used for is possibly the most important question to answer with AI. The more data collected, the more security issues arise. It seems that the infrastructure for the internet we have now is nowhere near adequate for the data of tomorrow. Yes, AI itself can lend a hand to security, and it’s beginning to look like it will, but without intentional security and the forethought to prevent or handle cyber attacks that don’t exist yet, it will still take a lot of learning to answer the question of AI security.
““If you think about security from the get go, then it changes the way everything is made,” Nate said.
Stay tuned for the next installment of 1 Billion Bits in November.