[Paper Review] Understanding artificial intelligence ethics and safety

David Leslie|arXiv (Cornell University)|Jun 11, 2019

Ethics and Social Impacts of AI276 references136 citations

TL;DR

This paper presents a guide identifying potential harms of AI in the public sector and proposes operational measures for governance, responsible innovation, and human-centered, interpretable AI.

ABSTRACT

A remarkable time of human promise has been ushered in by the convergence of the ever-expanding availability of big data, the soaring speed and stretch of cloud computing platforms, and the advancement of increasingly sophisticated machine learning algorithms. Innovations in AI are already leaving a mark on government by improving the provision of essential social goods and services from healthcare, education, and transportation to food supply, energy, and environmental management. These bounties are likely just the start. The prospect that progress in AI will help government to confront some of its most urgent challenges is exciting, but legitimate worries abound. As with any new and rapidly evolving technology, a steep learning curve means that mistakes and miscalculations will be made and that both unanticipated and harmful impacts will occur. This guide, written for department and delivery leads in the UK public sector and adopted by the British Government in its publication, 'Using AI in the Public Sector,' identifies the potential harms caused by AI systems and proposes concrete, operationalisable measures to counteract them. It stresses that public sector organisations can anticipate and prevent these potential harms by stewarding a culture of responsible innovation and by putting in place governance processes that support the design and implementation of ethical, fair, and safe AI systems. It also highlights the need for algorithmically supported outcomes to be interpretable by their users and made understandable to decision subjects in clear, non-technical, and accessible ways. Finally, it builds out a vision of human-centred and context-sensitive implementation that gives a central role to communication, evidence-based reasoning, situational awareness, and moral justifiability.

Motivation & Objective

• Identify potential harms caused by AI systems in the public sector.
• Propose concrete, operational measures to counteract harms.
• Promote governance processes and a culture of responsible innovation.
• Advocate interpretable, non-technical explanations of AI outcomes to decision subjects.
• Build a vision of human-centered and context-sensitive AI implementation.

Proposed method

• Review of the landscape of AI harms and governance needs in public sector contexts.
• Propose concrete, operational governance measures and responsible innovation practices.
• Emphasize the need for algorithmically supported outcomes to be interpretable by users.
• Advocate communication, evidence-based reasoning, situational awareness, and moral justifiability in AI deployment.

Experimental results

Research questions

• RQ1 What potential harms can AI systems cause in public sector settings?
• RQ2 What governance processes and practices can anticipate and prevent these harms?
• RQ3 How can AI outcomes be made interpretable and understandable to non-technical decision makers and subjects?
• RQ4 What characterizes a human-centered and context-sensitive implementation of AI in the public sector?

Key findings

• AI innovations can improve government provision of social goods but carry legitimate risks and potential harms.
• Governance and a culture of responsible innovation are essential to anticipate and prevent harms.
• Algorithmic outcomes should be interpretable and explainable to users in clear, non-technical terms.
• Implementation should be human-centered, context-sensitive, and grounded in communication, evidence-based reasoning, and situational awareness.
• Moral justifiability should be central to the design and deployment of AI systems in government contexts.