Immediate Prospects for AI and Robotics
Within the next ten years we are likely to see major advancements in the way robotics and autonomous systems (RAS) are used in areas such as manufacturing or packaging, leading to greater efficiency and generating cost savings.
A revolution is also under way in how these systems are used in labour-intensive industries such as mining and agriculture, where they are not just undertaking physical tasks but also interfacing with new AI-like technologies such as real-time sensing, intelligence and logistics, to make the processes safer and more efficient.
In other socially important areas that affect our day-to-day lives, such as transport or water and food security, RAS technology is increasingly applied to monitor, maintain and repair assets. There are also new novel capabilities where none existed previously such as innovative healthcare technologies, underwater mining or point-to-point drug delivery in remote geographical locations using drones.
Looking further forward, we could see the creation of micro-robotic systems designed to carry out repairs, deliver drugs or even repair cell structures within our bodies. At the other end of the spectrum there are massive, as-yet unexplored applications such as the large-scale 3D printing of structures for transport and housing. One of the key UK funding bodies, EPSRC, continues to support blue-sky research in these areas. One day they will become a reality.
Space is another frontier that has traditionally been very receptive to RAS technology. At the Edinburgh Centre for Robotics, we are collaborating with NASA and the Johnson Space Center (JSC) to develop humanoid robots capable of dexterous behaviour in complex environments for unmanned missions to Mars.
We are extremely privileged to be the only team in Europe with one of NASA's Valkyrie humanoid robots (funded partly through an EPSRC Capital grant). Valkyrie's human-like shape is designed to enable it to work alongside people, or carry out high-risk tasks in place of people. Edinburgh has a preferred partnership with NASA-JSC, making it the only institute outside of a couple of universities in the US with the capacity to address these challenges.
How do we support advances?
Dedicated centres for research and training excellence in the UK, initiated and funded by EPSRC, have been crucial in addressing major challenges for RAS, and have gone a long way to improving the visibility of UK robotics research outside of the country and Europe. It has helped attract lots of talent.
This commitment to long-term investment is very important for robotics, which is a high-investment high-return discipline requiring cutting-edge infrastructure, personnel and training. EPSRC's investments contrast with the scattergun approach we used to have, with robotics inevitably falling into the cracks between the ICT and engineering portfolios. The UK has evolved into a serious player in this domain in Europe and the trajectory is great. We have clearly put ourselves on the map.
In order to maintain this upward trajectory, there needs to be sustained investment in both people and centres. We need to encourage and provide incentives for investment from the large companies becoming interested in these centres, because it's a win-win situation. The key industrial players will develop and influence the next generation of technological leaders, while giving the centres the freedom and timescales to carry out disruptive, high risk innovations.
Questions to be answered
Advances in RAS and AI pose many social, ethical and moral questions. We need to ensure there is consultation not just with scientists and lawmakers but with wider society and stakeholders, the end users. We also need to manage the hype of what is possible and what is not.
Humans in general tend to overestimate the importance of new technology in the short term and underestimate it in the long term. This is very relevant to robotics because when people see an exciting new technology demonstrated, they think the problem is solved, and assume they will be getting a robot butler next.
Short-term overestimation of new robotics and AI technologies can lead to concerns about robots taking over the world and such like. I know first hand how hard it is for a bipedal robot to balance itself and stand up without falling over and breaking its ankle - we're a long way from having to worry about robot overlords. There is a gap between the vision of making robust, reactive systems and the challenges of realising it. Nevertheless, RAS technology will significantly and positively change the way we live, work and play in the very near future. I am confident of this.