RAIIMBY – Responsible Artificial Intelligence In My Back Yard, Here in North Saanich
by Jillian Buriak
World leaders in computer technology are leading the way to environmentally responsible, ethical AI
The headlines regarding artificial intelligence over the past 4 years were initially exciting, but now are increasingly alarming. As we all know, so much talent and passion resides in North Saanich, and the goal of North Saanich Currents has been to spread the word about the remarkable community we live in, and that includes two international leaders in the area of AI and computational technology. We were fortunate to catch up with Jonathan Schaeffer and Jennifer Griffin Schaeffer, to ask them about artificial intelligence as they see it, and the unique technology company they run from their office in Dean Park. Jon is a Professor Emeritus, and an internationally prominent researcher of AI and machine learning who developed the first computer program that could beat a human at any game (called Chinook, which played checkers). Jennifer is an expert in digital strategies for large educational institutions. Both are authors of several books in fiction and non-fiction, and have overseen the launch of at least 4 companies based on big data and computational technologies.
Jillian: Thank you so much for taking the time to talk to the residents of North Saanich about your insights into AI. Can you give us a 30-second definition of what an LLM is?
Jonathan: LLM is the technical name for what people generally think of as ChatGPT, Claude, or Gemini—these chat programs. Of course, they are much more enhanced in their capabilities beyond simple chat.
Jillian: How do they work?
Jennifer: They flip a coin!
Jonathan: What you need to understand is that people like me who work in the field of AI are magicians, like Harry Houdini, David Copperfield, or Penn and Teller. We create the illusion of intelligence, and just like magicians, what we achieve is all that matters, not how we achieve it. Penn and Teller don’t actually perform magic, but behind the scenes, they create the illusion of magic. It is the same thing with artificial intelligence.
My colleagues and I create the illusion of intelligence by coming up with ideas and algorithms that have no relationship to human understanding. So, a program like ChatGPT gives you incredible, thoughtful, knowledgeable, and in-depth answers, yet it has no understanding of what it writes back to you. What it does is ingest as much of the internet as it can reach and flip coins. If you type “How are ___”, it analyzes all of that data and finds that the most probable next word is “you” 50% of the time. Less frequently, it might find “your friends” or “the Schaeffers”. Whichever outcome the coin flip lands on is what it provides to you—with no underlying understanding, but successfully creating the illusion of intelligence.
Jillian: From when you started out in computer science, would you have predicted where we are now with AI?
Jonathan: I took my first course in 1977 and started my Master’s in 1979, so I have been doing this for nearly 50 years. The answer is no, I did not expect this to happen in my lifetime. It was baby steps forward for a long time, but on November 30, 2022, when ChatGPT was released, it blew me away. Ever since then, progress has been accelerating rapidly.
Jennifer: Jonathan has argued for a long time that it shouldn’t even be called “artificial intelligence,” but rather “augmented intelligence.” I have been around institutional strategy for years. One of the things I find strange since November 30, 2022, is how the entire world suddenly became a live research lab for an error-prone set of software. In all the decades prior, institutional strategy was entirely about mitigating risk—determining if something is cyber-secure, where data is stored, how it is transmitted, and how it is encrypted. Now, it feels like we threw all of that out the window.
Big enterprises might operate with locked-down LLMs internally, but the rest of us are driving down the highway at 100 miles an hour with no seatbelts, no car seats, and the doors taken off. That is how dramatically risk management has changed since 2022. It is why I am so supportive of what Jonathan started—building a business around private, secure AI. It can be done differently.
Jillian: Can you share an example or project where AI has made a very positive difference?
Jennifer: Take accessibility as an example. When someone posts an image or video, KIND AI can analyze it and automatically generate an incredibly detailed description for an alt tag. This gives visually impaired users a genuine understanding of what’s happening—such as three people talking on screen—rather than a generic, non-specific tag. A photographer we work with uses this daily to generate alt tags immediately, saving hours compared to manually writing detailed descriptions for hundreds of images. I now see what we do as NIMBY – Nature In My Back Yard, or kNowledge In My Backyard.
Jonathan: On a personal level, I have thousands of files on my laptop collected over many years—medical results, private financial information, family genealogy, and intellectual property. I want to analyze that data using AI without uploading it to Google or third parties. With Kind, I can query my own desktop data securely. It even quotes my own papers and notes back to me from decades ago that I had completely forgotten about, bringing my past work and ideas back to life. And no data centres involved since it runs on your laptop.
Jillian: Tell us a bit about your company and your local operations.
Jonathan: Our company is called Synsira (synsira.com). We relocated from Edmonton to North Saanich over two years ago, and our home office is based here in Dean Park. We are delighted to call North Saanich home for both ourselves and our business, and our goal is to build a major success on Vancouver Island that creates local tech jobs.
For those interested in learning more, Jonathan just published a non-technical book about AI: The Games that Computers (and Humans) Play: A Non-Technical Introduction to Artificial Intelligence. We will request that the library in Sidney acquire a few copies in both text and audio formats.
