AI week digest: reports on Ethics and Security

To support techUK’s AI week (23 April), we are bringing together the Observatory’s resources on AI in a series of themed digest pieces. This article brings together all the recent reports focusing on the ethical and security issues relating to AI.

AI in the UK: ready, willing and able?

Published, 16 April 2018

The House of Lords Artificial Intelligence Committee publishes a report on principles for AI development and adoption, which include taking precautions to reduce prejudices being built into automated systems and maintaining open data protection structures.

The report includes recommendations such as investing in skills training for people whose jobs will be displaced by AI and investigating the need for new liability laws to cover harm caused by AI systems.

Link to report

Algorithmic Impact Assessments: A Practical Framework for Public Agency Accountability

Published, 08 April 2018

The AI Now institute has outlined their development of an Algorithmic Impact Assessment (AIA) framework. This framework was developed to support communities and stakeholders to assess the claims made about autonomous decision systems and whether their use is appropriate in a particular setting, or at all.

Link to report

Artificial Intelligence: Emerging opportunities, challenges and implications

Published, 28 March 2018

This U.S. Government Accountability Office report recounts the outcomes of a forum on how artificial intelligence in high-consequence sectors.

This report explores the opportunities and challenges that artificial intelligence presents for cybersecurity, automated vehicles, criminal justice and financial services.

Link to report

The Malicious Use of Artificial Intelligence: Forecasting, Prevention and Mitigation

Published, 21 February 2018

This report from the Future of Humanity Institute reviews the security threats artificial intelligence technologies could pose if used for malicious purposes.

Example scenarios are presented to demonstrate how threats may change within the fields of digital security, physical security and political security along with four key recommendations to mitigate against such threats.

Link to report

The Future Computed: Artificial Intelligence and its role in society

Published, 17 January 2018

In this book, Microsoft offers its view on the direction of AI technology and its potential societal impacts.

The book identifies fairness, reliability, privacy and security, inclusivity, transparency, and accountability as six principles that should guide the use and development of AI. For the full potential of AI to be realised businesses, academics, governments and society must work together to create AI systems that are trustworthy.

Link to report

Ethically aligned design: A vision for prioritising human well-being with autonomous and intelligent systems

Published, 12 December 2017

This IEEE Standards Association document outlines recommendations on the ethics of autonomous and intelligent systems.

This document is an update of the original EAD, which stems from a global consultation of 250 cross-disciplinary thought leaders and produced pragmatic recommendations for technologists, policymakers and academics.

Link to report

Optimism and ethics: an AI reality check

Published, 7 November 2017

This report from Sage finds that a majority of UK and US consumers reject extremely negative predictions of the future of AI. Instead, the report argues that the more pressing concern is how to make the adoption of AI ethical.

The report recommends that AI creators develop more consumer-friendly security guidelines and highlights that there is a need for the tech industry to educate consumers on the potential benefits of AI, as well as ensure that solutions meet these ideals.

Link to report

AI 2020: The Future of Defence and Security

Published, 27 November 2017

This report from the Artificial intelligence & Intelligent Automation Network summarises a survey of 100 defence and security specialists on how AI and intelligent enterprise technologies are implemented in businesses and what the impacts of this could be by 2020.

Barriers to intelligent enterprise technology adoption are highlighted, as well as potential methods of overcoming these.

Link to report

Harnessing Artificial Intelligence for the Earth

Published, 19 January 2018

This report from PwC, in collaboration with the World Economic Forum and the Stanford Woods Institute for the Environment, explores the potential opportunities and threats that AI technology poses to the environment.

It argues that the opportunity for AI to be harnessed to benefit earth is substantial and recommends to deliver "responsible AI", collaborate for interdisciplinary solutions, and direct finance for innovation.

Link to report