Why are we doing this?
Some algorithms are able to learn by themselves, finding patterns in big data sets and predicting the outcome of new events. Autonomous systems are being built to take on tasks independently.
Governments, companies and public services are increasingly using algorithms to help them make decisions. There is hope that machines could help solve some of the most difficult of humanity’s issues. But algorithms are not free of the problems that plague human decision-making, and they even add some more problems of their own.
Nesta hopes to highlight the opportunities of creating automated systems as well as the risks of reducing our complex world to a series of computable steps. By engaging directly with those at the cutting edge of machine-supported decisions, Nesta hopes to become a champion of best practices that are both responsible and profitable.
What are we doing?
Nesta’s research in this area started with commentary on examples of automated systems like self-driving cars or drones. It also includes some work on automation and employment, particularly as it affects creative occupations and the future of employment in the creative economy.
Machine learning processes allow higher levels of automation and processes free from human intervention. Several of our Hot Topics events have covered the use of machine learning algorithms, particularly the Sounding the Cyber Security Alarm and Machines that learn in the Wild in 2015. The Machines that learn in the Wild report looked at different kinds of learning algorithm and their uses so far.
As more and more organisations invest in or investigate these opportunities, Nesta is also looking at the ramifications of widespread algorithmic reasoning. The UK Government is already thinking about how it can support the ethical use of data science internally. But there are still a lot of unanswered questions as machines continue to make more important decisions. Understanding who is responsible when something goes wrong or what the societal impact of these decisions will be are hugely complex issues. The shadow of the smart machine - a workshop and series of blogs in 2016 - kicked off Nesta’s work looking at these issues.