Intelligence Augmentation: Machine Learning For The Real World
In part two of this series, John Kane discussed the benefits of IA (aka Intelligence Augmentation or Augmented Intelligence) technology, including how it can enhance employees’ abilities to be more emotionally intelligent and perceptive. This third and final part of John’s series looks at how IA is impacting the broader society and day-to-day life outside of the workplace.
The past decade or so has been revolutionary for machine learning, in particular since Geoffrey Hinton and colleagues demonstrated in 2006 how to efficiently train deep neural network models. This breakthrough powers so much of the technology now ubiquitous in our daily lives. Artificial Intelligence (AI) systems have started gaining significant adoption both in large-scale enterprise where there is the potential to deliver significant improvements in efficiency and quality of services, but also in people’s everyday lives.
Hinton’s breakthrough fortunately coincided with the rapidly increasing availability of large volumes of data and developments in high performance computing, which has enabled the evolution of novel new product offerings which can be deployed at scale. Some of the most prominent examples of AI are currently in the personal voice assistants like Siri and Alexa. Going a step further, the Google Duplex system demonstrated the potential of the new capabilities in machine learning and AI, with a voice assistant which can automatically carry out certain tasks for you, e.g., booking appointments. At the same time, however, this has built somewhat unrealistic expectations in the minds of the general public. Google Duplex, impressive as it is, currently works just for limited domains. There are many challenging outstanding machine learning and language technology research obstacles to be overcome before a system like Google Duplex to increase the scope in which it can operate.
Adoption of machine learning-powered systems can more rapidly increase if they are designed in a user-oriented way where the goal is to empower the human users rather than to replace them. This is why it is so important for these technologies to be developed as a result of collaboration not just between software engineers and machine learning scientists, but also behavioral scientists and human computer interaction experts. Modern automated systems are really good at being repeatable, consistent and objective, and at solving complex narrowly defined problems. Where they fall down is with problems with a broader scope and for situations where a social or interactive communication is needed. Contrastingly people can deal with out-of-scope issues more gracefully and can provide much more engaging and meaningful interactions with other people. But at the same time we find it hard to be consistent and objective, particularly under stressors like fatigue. By using an intelligence augmentation (IA) approach, where the technology seeks to empower and complement people’s innate skills, a huge amount of value can be brought both to people’s personal lives and to the enterprise, with the machine learning capabilities that exist today.
We have already seen IA technology produce significant value in day to day life, for example:
- Trackers (e.g., FitBit, Apple Health) provide objective health and fitness related data back to users so that they can be more self-aware of their health and activity state. This objective self-awareness has already been demonstrated to result in actions which lead to health improvements. A similar self-awareness, but this time of behavioral and mental health related indicators, can be enabled using the app from CompanionMx, which provides objective feedback to people on their social connectedness, energy levels and overall mood.
- Despite the current media hype about the future potentials of fully automated driving, augmenting and empowering manual drivers is happening already today. Modern cars are using machine learning based technology to help drivers stay in their lane and to be more aware of obstacles and other cars around them. Others, like MIT spin-out Affectiva, have developed technology which detects driver distraction and fatigue and enables mitigation strategies to avoid crashes.
While riding the peak of the technology hype cycle, the promises of artificial intelligence (AI) have captured the public’s imagination about futuristic technologies previously only found in science fiction novels — while at the same time alarming them about ethical concerns. These are, however, elevated expectations and significant scientific hurdles remain to meet those promises. But as we descend from this peak and these deficits become more apparent, the potential of empowering and enhancing human behavior with a human-oriented augmented intelligence solution maintains tremendous potential to deliver value in people’s personal lives and in the workplace. The exciting part is that we do not have to wait for the machine learning capabilities to achieve this — they are already here!