The institute aiming to keep AI and data in check
According to programme manager Reema Patel, the Ada Lovelace Institute is an organisation looking at the societal and ethical implications of data and AI, with a mission "to ensure that data and AI work for people and society.”
Patel explained that one of the ways in which the Institute works with partner organisations is to facilitate workshops and public events, particularly with techUK and the Omidyar Network. The attendees at one workshop held in collaboration with techUK explored the roles and responsibilities for industry when embedding ethical AI. They put forward ideas that could help the industry deal with some ethical issues.
These included promoting cultures of inclusion and diversity so that tech companies engender open-mindedness, and increasing consumer engagement and user involvement and control over decision-making. They also put forward the idea of external accountability systems, for example ethical audits that would measure and enforce ethical practice, as well as the creation of interdisciplinary and multi-practice on ethics and how ethical practice would look.
"Fellowships second people to corporate and private sector organisations using AI."
She also mentioned that another tangible action the Institute is taking at the moment is the development of Ada Lovelace Institute fellowships. “These give people with ethical background training that are very committed, the chance to learn from the Ada Lovelace Institute, but also to be seconded in corporate and public sector organisations that use AI in order to apply what they’ve learnt, but also to understand what some of the structural challenges might be and to feed that back.”
The Ada Lovelace Institute identified a gap in organisations looking at how data and AI could “help inform and shape a better world.” She mentioned the good work being done by investigative journalists in reporting on particular instances of the way in which data and AI is impacting different groups in society. An example of this is work by ProPublica on uncovering the flaws in the Compas recidivism risk algorithm.
"We're in an era where trust and legitimacy have been eroded."
Patel added that her organisation lacked the tools and search mechanisms to track such issues, but also said that collaboration is a key part of the institute’s work going forward. “We are very keen to reach out to networks in order to build a better understanding and a better picture of how that impact might be measured.”
She put the need for the organisation into context by saying that, while many say that we are currently living through the "Fourth Industrial Revolution", one might suggest that we are in fact experiencing the "Fourth Information Revolution". "We are experiencing challenging technological times, we’re in an era where trust and legitimacy has been eroded.”
"We've got further to go to ensure data and AI work for people and society."
The Cambridge Analytica and Facebook scandal alongside the first pedestrian fatality of a self-driving car this year have “illustrated the extent and the way we’ve got further to go to ensure that data and AI work for people and society.”
The Ada Lovelace Institute was launched as a partnership by the Nuffield Foundation in late 2017 with a two-part mission to examine the impact of data, algorithms and artificial intelligence on people and society, and to promote and support ethical practice in the public interest. The former CEO of the Alan Turing Institute, Sir Alan Wilson, was appointed executive chair and the organisation is still recruiting for a board of directors.
Named after the woman heralded as the world’s first computer programmer, its partner organisations are The Alan Turing Institute, The Royal Statistical Society, the Nuffield Council on Bioethics, the Wellcome Trust, The Royal Society, The British Academy, techUK, and the Omidyar Network. The Nuffield Foundation has committed £5 million over five years for the Ada Lovelace Institute.
Reema Patel was speaking at the National Centre for Research Methods event ,"Social science methods and automated data algorithms".