AI and Humans: Focusing on what we can achieve together
Dystopian fever spiked again last week as PwC issued a report predicting UK jobs carnage at the hands of robots over the coming decade. "Artificial Intelligence" has dark overtones in the media and the popular imagination. Some of us are buzzed about this new era of augmented human performance and meaningful human computer interactions. Others worry for their livelihoods, even their personal freedoms.
AI technology has improved significantly over the last few years, mainly because of a combination of a huge increase in the availability of data and cheap computation. This means that AI researchers now have meaningful amounts of "training data" to train their models and the computational power needed to run complex learning approaches. The sudden acceleration of AI performance is a thrilling development for data scientists, and great seed material for ominous sci-fi and gloomy consultancy forecasts.
Rather than viewing increasingly sophisticated AI as a threat to human intelligence, jobs, even freedom, there is another view of the future that sees us unlocking a new wave of innovation by building systems that allow humans and machines to work together, each playing to its own strengths.
There are already examples of these types of "collaborative AI" solutions that help to amplify rather than replace human intelligence. After IBM’s Deep Blue beat Kasparov in chess in the 90s, the chess community started to embrace AI as a means to improve human chess ability. Now competitions that accommodate human+machine participants are beginning to emerge.
Elsewhere, data mining and machine learning systems have helped scientists to shift through huge volumes of existing data to discover new drug therapies and side-effects. This is possible because humans and machines are working together rather than in competition. Vast new terrains of knowledge are opening up to us, but we can only go there hand-in-hand with AI.
As we become more and more used to "digital assistants" that help us to book concert tickets, schedule appointments and generally run our lives, so we may become accustomed to "digital collaborators" that help us to perform at new levels of intellectual ability.
There’s only one way this is going to work. We need to focus not just on the AI research but critically on the complementary "interface and interaction" research that will provide an intuitive communication and collaboration platform to support natural and intuitive engagement with these digital collaborators. The experience of "collaborating" with a machine needs to feel fluid, human. The machine must respond not only to our commands but to our emotions, physicality, creativity. Happily research and development in this field is gaining momentum in tandem with AI.
The Insight Center for Data Analytics, Ireland's national data research institute has over 450 researchers working across and beyond data disciplines: it’s this marriage of data science and fields such as physiotherapy, chemistry, materials technology and psychology that is offering the greatest promise in developing meaningful communication between humans and machines.
Our work in the Insight Center is enriching the exchange between artificial and human intelligence across a whole branch system: chemical signals from the skin and eye, electrical signals from the brain, body language, facial expression recognition, personal communications and digital activity and gross and fine motor movement.
We might imagine a combination of the next-generation touch-based interfaces like the ones used by Tom Cruise’s character in Minority Report, combined with more sophisticated voice and vision based interfaces that will allow us to more easily describe our needs and respond to algorithmic suggestions.
For example, emotion analysis software currently in development will allow, in the near future, for cars that can "read" your tired and emotional state, and take your e-keys off you. A "digital conscience" of this kind could save you from the career suicide of sending that furious mail before you’ve slept on it. A computer that can discern sentiment is exactly what social media providers need right now in the hunt for abusive content. This is where computers assist us, make us better at being human, rather than eclipsing us or leaving us with nothing valuable to do.
The implications in the physical realm are very exciting. Athletes have pushed incrementally ahead, generation after generation, in the quest to keep overturning human records. Performance enhancing drugs have taken this endeavor in a dark direction, but AI can redirect the effort. Human/AI symbiosis in sport and exercise has enormous potential for injury prevention and performance optimization. The kind of gait, glucose, sodium and vital signs monitoring that was once only possible under sophisticated lab conditions will soon be available to the ordinary individual training for a 10k. This democratization and personalization of physical monitoring will give not just athletes but everyone more autonomy over their health management.
In terms of processing information, let’s face it, our human capacity to sift through available information was long ago left in the dust. Our unassisted brains will never be up to the task of making meaningful use of the data firehose blasting information out of everything from intelligent fridges to plush toy nursery cams. Working with AI, we can use our immense powers of problem-solving and invention to actually find the meaningful patterns in the data. The new human activities likely to spring from this alone are limitless -- for every traditional job that passes into history, a new role, engendered by AI, will likely spring up in its place.
This PwC report has prompted a predictable slew of scary headlines, but read the report itself and you’ll find a more balanced outlook. Its authors are chirpy about UK employment in the Age of the Robots, predicting higher value, more creative jobs, boosted productivity and additional roles in less "automatable" parts of the economy. There are certain to be jobs casualties, as there have been after every significant technological disruption since the Industrial Revolution. However, the potential transformation of human work offered by AI collaboration is a very exciting prospect.
Published under license from ITProPortal.com, a Future plc Publication. All rights reserved.