AUSTIN, TEXAS -- I'm attending my first SXSWi conference and in true fashion, it is living up to its hype -- even, as it turns out, about my own future.
I attended a discussion on Big Data and AI: Sci-Fi vs. Everyday Applications. The panelists included Robin Goad, Rayid Ghani, Dr. Doug Lenat and Kris Hammond — all experts in Big Data and AI. To kick off the discussion, Robin asked the experts, “What you believe we are doing today with the combination of Big Data and AI?”
In response, Kris asked the audience, “How many of you have looked at a spreadsheet on a daily basis?” As I’m sure you would guess, almost every hand in the room was raised. Kris elaborated on the concept that AI can be a tool we use to complement our everyday life. He explained how in this day and age we have massive amounts of data available. We need methods to extract the data for everyday applications. Big AI enables efficiencies of data analysis that would take humans weeks, in a matter of minutes. People like Kris are working to increase workplace efficiencies by using the available data in a smarter method. “Analytics, inference and communication make up an exquisite approach to using data in our world,” said Kris.
To me, the most interesting part of the discussion was around integrating AI into daily life and the relationship it will create. Dr. Lenat made an important distinction between trusting AI to vacuum your rug vs. taking care of your baby. “This gap has two parts,” said Dr. Lenat, “first being common sense and the second is the ability of AI to introspect of its own line of reason to be trustworthy. You need both types of AI for real, functional, reliable, safe intelligence.” What we want AI to do, and what we trust the AI to do doesn’t always go hand in hand. As Dr. Lenat stated, “AI should be our partners to keep our minds flexible.”
Dr. Lenat shared a personal experience on the capabilities of AI. When his daughter was two years old, she was diagnosed with meningitis. After weeks of doctor visits, no one could tell them why. His company was working on a software technology to expedite medical diagnosis, and it came up with the same diagnosis and treatment plus an explanation as to why. This type of software technology digests massive amounts of data on any symptom and the corresponding disease with structured knowledge and inference to output real answers. The diagnosis proved to be reliable. Technicians might trust the AI; however, patients do not. This example of AI isn’t meant to replace the role of a doctor. These types of efficiencies would help health care providers automate diagnosing patients so that they can use their time and resources elsewhere.
In a separate but related session, Jerry Kaplan had a more realistic view on AI, what it is and what it is not. He started the presentation with explaining the difference between thinking and computing. The human brain will process information differently compared to a computer, especially when it comes to things like caring for a child or translating another language. We can specify the common sense, context for the known and unknown knowledge, and basic inferences, but you can’t code human semantics.
Artificial Intelligence will continue to evolve and progress. I’m excited to see how these software technologies integrate into everyday life, increase productivity and if they become our partners.