‘Autonomous’ technologies can free human beings from repetitive tasks, increase productivity and even look after us when we’re old and frail. But with robot carers appearing at our bedsides and driverless vehicles arriving on our roads, what are the safety implications for human beings? Lloyd’s Register Foundation’s partnership with the University of York in the Assuring Autonomy International Programme (AAIP) is helping ensure these new technologies help us rather than harm us.
With the AAIP, we’re working with an international community of developers, regulators,
researchers and others. Together, we’re working to ensure that the public can benefit from the safe, assured and regulated introduction and adoption of robotics and autonomous systems.
The AAIP’s work particularly looks at the points where intelligent machines and fragile humans come together – to ensure that the new technologies become part of a brighter, more connected future rather than the dystopian visions of science fiction.
We’re working with the AAIP in a number of ways – firstly, to develop a framework for assuring the safety of autonomous technology. We’re also supporting their training and education work, and the development of a Body of Knowledge – a global knowledge bank - to help the world make the most of machine learning, robotics and artificial intelligence.
Dr Richard Hawkins, Senior Research Fellow, said: “The funding from LRF has enabled us to create the Body of Knowledge resource that benefits the whole international community involved in the development, introduction, adoption, and regulation of autonomous systems in any sector or domain.”
Last year, our grant to AAIP helped develop a safety assurance process, AMLAS (Assurance of Machine Learning for use in Autonomous Systems). This generates assurance evidence at every stage of the engineering of machine learning components. It’s the first systematic,
documented approach to safety assurance of machine learning components, with the aim of giving others the confidence they need to use, certify or regulate the component or the system it is part of. It shows that machine learning components can perform their tasks safely, with the lowest practicable risk of harm to humans. The AAIP team are currently validating the AMLAS process through their work with industrial partners.
Another key part of AAIP’s work is collaborating with industry on demonstrator projects in areas where robots, machine learning and other autonomous technologies have the potential to improve our lives. This includes a semi-automated quarry, in collaboration with Volvo Construction Equipment, where they are investigating and adapting hazard analysis techniques.
Another demonstrator project is helping to ensure that the use of ‘assistive technologies’ in the home is only done safely. Robots can provide support for a range of care-related tasks such as physical and social assistance, physiotherapy and rehabilitation. But so that they are only used to enhance the lives of those who need care, rather than adding risk, AAIP is funding a demonstrator project that is investigating and evaluating the safety and regulatory requirements of assistive robots that come into close contact with humans. The project team have developed a prototype modular overhead/ceiling robotic system that can provide physical and cognitive
assistance to frail older adults.
As part of the project the team are experimenting with a range of practical uses, gathered from potential end-users, health professionals, carers and regulators, as well as potential commercial manufacturing partners. They are looking at functionality but also the human factors – trust, attention, perception and learning - associated with human-robot interaction (HRI).
Our grant to AAIP is also helping them build up an international community of experts. AAIP fellowships have attracted experts from all over the world. The outputs from all areas of AAIP’s work feed into the growing AAIP Body of Knowledge, an online, curated resource that is freely accessible. This provides practical guidance for anyone involved in developing or regulating autonomous systems.
By funding AAIP’s work to advance the safety of robotics and autonomous systems, we’re
supporting work that will benefit all of us into the future.