Companies that want to benefit from artificial intelligence should be looking at top-line value versus cost savings and how they can augment existing processes, according to the chief strategist of IBM’s Watson division.
In his keynote speech at the Canadian Innovation Exchange (CIX) in Toronto Tuesday morning, Marc Teerlink predicted that in five years every major decision — from comparing contracts in a business to approving a mortgage — will involve consulting an AI system of some kind. Before we reach that point, however, companies need to make sure they have the right information in place for AI tools to learn what they need to assist with various tasks.
“I always ask businesses, ‘Where are you going to get your training data?’ (AI) is not telepathic, contrary to what you might think,” he said.
An AI tool might use an image of a banana to know identify the fruit, for example but might need what Teerlink called a “custom classifier” to distinguish between a ripe one and one that’s gone bad. These distinctions are critical in many business processes, he said.
Teerlink said IBM Watson is being used by health-care providers such as Memorial Sloan Kettering to accelerate the process by which cancer patients are screened for clinical trials. Approximately 20% of patients are eligible today, but less than five per cent are matched because of the time physicians need to make a determination. Teerlink said Watson augments the physicians’ knowledge by going through a series of parameters such as the patient’s genetic disposition, along with an analysis of more than 15,000 studies, to provide a recommendation. This reduces the time a physician needs to as little as 14.5 minutes to digest.
“This is not replacing physicians but augmenting their knowledge,” he said. “Let the robots process and the humans think.”
Besides IBM’s Watson, many of the firms working in the AI space are trying to become more sophisticated in finding the right target market and helping to explain where the technology could help.
Vancouver-based 1QBit, for example, is developing products and services that will bring AI capabilities to the quantum computing sector. According to Andrew Fursman, the firm’s co-founder and CEO, what seems like a theoretical area to most of the world today will benefit in many ways with AI behind it.
“We were looking at areas like computational finance, energy. Things that could generate millions of dollars,” he said.
Clarifai, based in New York, uses machine learning to assist organizations in organizing photos, videos and other digital assets. CEO Matthew Zeiler said he originally created a consumer app to do this with his own images, but is now selling that off as his company focuses on bringing the technology to businesses.
“Marketing to consumers and developers and enterprises is completely different,” he said, adding that while initially some firms were put off by AI, such tools are now at the forefront of their thinking. That kind of worry has gone away. It’s very easy to have conversations and talk about our customers.
Kindred.ai, which is creating AI to make robots act more like human beings, has been on a similar journey. According to founder Suzane Gildert, the company had originally targeted the home market, before recognizing that consumers couldn’t really afford it.
“The price point was really hard,” she said, which mean pivoting to firms in the e-commerce or industrial equipment space. “Customers in that industry okay with buying large pieces of equipment and are happy with a small increase in output.”
CIX continues on Wednesday.