The AI in use today is actually a group of related technologies, including machine learning, supervised learning, and computer vision that allows companies to create automated tasks on a large scale. Supervised learning is the “backbone” of most artificial intelligence systems in use today.
“There are a lot of real risks to AI,” said CNET senior editor Stephen Shankland, like “job displacement, more advanced weapons, and new ways for humans to be bad to other humans.” But sci-fi scenarios “vastly overstate the risk.”
We have AI in our homes and workplace, like Amazon’s Alexa, Google Home, and Apple’s Siri. These devices are not programmed to be independently intelligent, but to excel at so-called “smart” interactions. Basically, they are designed to be guides or assistants and provide memory extension.
Supervised learning and AI
With artificial intelligence, automated interactions are powered by algorithms that are coded by humans. In supervised learning, the training data consist of a set of training examples. Each example is a pair consisting of an input object (typically a vector) and the desired output value (also called the supervisory signal). What does this all mean?
The machine learns by processing massive amounts of data and in the process, AI learns to adapt to ever-changing real-world conditions. For example, A supervised learning algorithm analyzes the training data and produces an inferred function, which can be used for mapping new examples (machine learning).
Supervised learning has played a key role in the advancement of self-driving cars. “A self-driving car knows how to avoid a pedestrian and knows how to stay in a lane because humans have sat down with images taken from the front of these cars and said, ‘this is a human, this is a lane that marks the edge of the road,'” said Figure Eight CTO Rob Monroe.
“It’s these thousands and thousands of hours of human feedback that ultimately enables a self-driving car to be trained to automate the task” of driving.
Other industries using AI
Using “computer vision,” AI has enabled the breakthrough of a number of technologies in the healthcare field. Arterys is an online medical imaging platform. Fabian Becker, the CEO of Arterys says using AI to assist in making a diagnosis can save many lives.
“We are working on automating that [process], and it’s really a quantum leap because the system can propose a suggestion to the physicians” in a fraction of the time it takes a human. The Artery platform is entirely web-based and cleared by the FDA.
The world of agriculture is also benefiting from machine learning and computer vision AI systems. Blue River is just one of many new companies using computer vision and AI to make it easier for machines to identify weeds, select fields and optimize planting times.
Blue River has introduced an advanced “See and Spray” machine to make farming even smarter. “[Our database] is now close to a million images, across many different farms in many different geographies,” said Jorge Heraud, CEO of Blue River Technology.
The company’s AI-enabled machine is easy to use and runs at speeds up to 12 miles per hour, and can be towed behind a tractor. “It looks at every single plant 50 times per second” and determines “‘is this a crop or is this a weed,’ then applies herbicides only to the weed and nowhere else.”
Artificial intelligence is easily misunderstood, said quantitative futurist Amy Webb, because AI is tied to the way our brains work and how we think. It really can’t be any other way because AI is tied to what we as humans can teach it, based on our own input and learning experiences.
Which means, Webb said, we’re all using AI all the time. “It’s part of our everyday lives.”
