5 industries ripe for human-machine learning

Machine learning has been a constant on tech trend lists for years. This year, it’s time to embrace what humans can learn by interacting with machine learning.

As Google’s head of Machine Intelligence, Blaise Aguera y Arcas noted in a recent Medium article: “Machine intelligence will expand our understanding of both external reality and our perceptual and cognitive processes.”

In the spring of 2016, Google’s AlphaGo software, fueled by machine learning, beat the world’s greatest human Go player, Lee Sedol. The victory was a major milestone for a specific type of AI, called deep neural networks, that is more closely modeled on the way humans think.

The AlphaGo team refined the machine’s Go-playing prowess by training it on 30 million moves from prior games, and also by pitting AlphaGo against human experts. While machine learning was the clear protagonist of the story, a funny thing happened during these human-machine face-offs. In training and playing against AlphaGo, the human Go players also improved. While the use of neural networks in AlphaGo was proof of how human thinking influences machine learning, AlphaGo’s interaction with human players also suggested a future in which machine learning could influence human thinking. We are already indirectly learning from machine learning in other ways, whether by refining our music tastes while we help Spotify refine its algorithm or by learning about the brain by observing neural networks learning.

What happens when we approach machine learning not as a replacement for human expertise, but as a partner in a collaborative relationship where humans and machines learn from each other? Could observing a computer make new connections between words make us more creative writers? Could we teach someone a new language and refine a computer’s translation abilities at the same time? Everyone is talking about machine learning. Let’s talk about human-machine learning, as well.

Learning from machine learning could have an immediate impact on a number of industries. Below are five predictions for how human-machine learning could impact our lives in the coming years.

1. Education

Education is one of the areas with the clearest opportunity for embracing human-machine learning. For example, adaptive learning draws on machine learning to help tailor and evolve educational experiences based on a student’s learning style. Companies from education stalwart Pearson to start-up DuoLingo are embracing it, and the software — at least in the case of DuoLingo — also refines its translations over time as it draws on human input. As online and blended learning continues to grow, organizations that can creatively embrace the reciprocal relationship between humans and machines could have a competitive edge. They could also help redefine what it means to learn — for humans and machines.

What could this look like?

The Magic School Bus: Students have personalized learning plans that adjust to their behaviors (like the Nest thermostat) and give recommendations for new content (like the Netflix dashboard). How frequently do you need to see that Flashcard? Are you a visual learner? Do you learn better in small groups? Adaptive learning platforms could create collaboration between students, educators, and technology.

2. Human capital

Human capital, from recruiting to management, offers an ideal context in which to embrace the reciprocal relationship between human and machine learning. A number of startups, including Belong and Prophecy Sciences, are exploring machine learning as a way to augment the hiring process. Google’s People Operations team, and others, have pioneered the use of data-driven human capital. We could enhance professional development by using machine learning to identify and predict human capital trends and needs and then create a dialogue between employees and algorithms.

What could this look like?

Human-Machine Resources: Employees get assigned a human manager — and a machine learning coach — to help them develop throughout their career.

3. Venture capital

VC firms’ investments in artificial intelligence-related startups have been growing steadily over the past five years, but the opportunity to draw on machine learning to drive VC firm investments remains largely untapped. Venture capital, with its combination of interpersonal relationships, insider knowledge, and instinct balanced by quantitative trend identification and analysis, could be an ideal context for human and machine learning collaboration.

What could this look like?

AI Combinator: A startup incubator driven by VC expertise and machine learning, drawing from the latest angel investing and industry trends to identify new market opportunities. By interacting with machine learning, venture capitalists could develop new investment strategies or targets they may otherwise have missed.

4. Psychology and behavioral science

A new MIT study suggests that an algorithm can predict human behavior more quickly and more reliably than people can. As machine learning evolves, it has the potential to help us gain further insight into how we think and behave and can motivate us to change those behaviors when we want to. Whether through therapy, building daily fitness habits, or encouraging retirement investing, behavioral interventions across industries offer a wealth of opportunity for human-machine learning.

What could this look like?

HaBit: A Fitbit for habit building that helps people track their behaviors and provides personalized motivation and feedback to enable behavior change when they need it most.

5. The arts

Perhaps the least practical and most open-ended human-machine learning could change the way we approach the creative process. Machine learning is not just analytical, it’s also generative. It can identify existing patterns (e.g. a cat versus a blueberry muffin), but it can also generate new content, whether visual images or musical compositions. GoogleBrain, the team recently featured for overhauling Google’s approach to translation, and AI more broadly, has launched Magenta to determine whether we can “use machine learning to create compelling art and music.” The implicit follow-up question is: How might we collaborate with the creative output of machine learning? And, in the process, how can we learn from it and evolve our own creative process?

What could this look like?

Co-creation: Artworks are co-signed by an artist and an algorithm. Musicians, writers, and artists see machine learning as a collaborative partner and influence, and in turn actually create differently.

At this point, we know that machine learning will impact industries and the nature of work. But as it replaces parts of our daily lives, how might our collaboration with machines also influence us as humans — how we think, learn, and create? We are looking at a future in which humans and machine learning could be collaborative partners, whether that’s in a classroom, on a canvas, or in a boardroom.