Digital Catapult launches Ethics Framework for AI and machine learning

Innovation centre Digital Catapult has unveiled its first Ethics Framework, which it hopes will increase the adoption of ethical standards by organisations developing artificial intelligence (AI) and machine learning technologies, especially startups.

Digital Catapult said the Ethics Framework is focusing on startups because they are a good testing ground for ethical tools. Unencumbered by legacy infrastructures, startups can more readily benefit from adopting ethical practices from day one, it said.

Luciano Floridi, chairman of Digital Catapult’s machine learning garage ethics committee, the independent body responsible for developing the framework, said: “It is critical that businesses of all sizes start to think and act ethically and responsibly.

“We know from working directly with early-stage AI companies that a highly practical framework, which speaks in the language and caters to the needs of those developing AI-enabled products or services, is required at this crucial stage in the formation of the UK’s already remarkable AI community.”

The Ethics Framework urges companies to: be clear about the benefits of their product or service; know and manage their risks; use data responsibly; be worthy of trust; promote diversity, equality, and inclusion; be open and understandable in communications; and to consider their business models.

These points are further explained for organisations by Digital Catapult. For example, “use data responsibly” includes questions on data anonymisation, whether the data usage can be clearly communicated, and whether there are processes in place for discovering and dealing with inconsistencies in the data – all of which can affect how ethically data is being used.

However, the framework is only designed to guide the application of ethical development practices and support companies as they try to adopt these, rather than establish a strict set of rules for organisations to follow.

Digital Catapult CEO Jeremy Silver said: “At this very early stage in its development, presenting the industry with a set of enforceable rules might deter companies from trying to navigate and understand what responsible AI looks like in practice. The Ethics Framework has been designed to provide guidance for companies to explore what role ethics might play in their projects and to help share best practice across the industry.”

Further guidance on how to navigate the framework is provided by Digital Catapult’s ethics committee, which comprises a steering group and a working group. The former will oversee the creation of the principles and tools required to promote responsible AI development, while the latter will work directly with startups through Digital Catapult’s machine intelligence garage programme, which helps businesses to access the computation power and expertise they need to get to grips with intelligent systems.