• The role of chief AI ethics officer (CAIEO) is on the rise at leading enterprises as digital transformation becomes more complex and AI adoption grows rapidly across industries;
  • Forward-looking companies are turning to the CAIEO role to put into operation corporate values related to AI across the organization's divisions. CAIEOs need to ensure that the AI technology being developed, used and deployed is trustworthy; and that developers have the right tools, education, and training to easily embed these properties in what they produce;
  • CAIEOs should have multi-disciplinary knowledge of AI techniques, tools and platforms, AI risks and its impact on society, business strategy, industries and public policies, as well as good communication skills;
  • A new report by the Global Future Council on AI for Humanity explores educating the CAIEO role and others to put into operation AI fairness across an organization.

Artificial intelligence (AI) affects the lives of billions of people, rapidly transforming our society and challenging what it means to be human. Some think AI is just a buzzword, but it’s powerful enough to enable solutions in every sector, from personal digital assistants to fraud and failure prediction, self and assisted driving, and health diagnostics. AI can help personalize education and tutoring, create new jobs and assist in tackling the COVID-19 pandemic and its aftermath.

Alongside its positive effects, some AI applications raise legitimate concerns and risks. AI ethics is the multi-disciplinary and multistakeholder field of study that aims to define and implement technical and non-technical solutions to address these concerns and mitigate the risks.

AI solutions could, for example, unintentionally generate discriminatory outcomes because the underlying data is skewed towards a particular population segment. This could deepen existing structural injustices, skew power balances further, threaten human rights and limit access to resources and information.

Some AI systems could behave like black boxes with little or no explanation of why they make their decisions. According to FICO's latest report dedicated to the State of Responsible AI, two-thirds (65%) of respondent companies can’t explain how specific AI-based decisions or predictions are made. This could erode trust in AI and thus hamper its adoption, reducing the positive impacts of this technology. It could also damage a company’s reputation and the trust of its clients, as well as contradicting company values.

The main goal of a Chief AI Ethics Officer (CAIEO) is to make AI ethics principles part of operations within a company, organization or institution. A CAIEO advises and builds accountability frameworks for CEOs and boards on the unintended risks posed by AI to the organization. They should help companies comply with existing or expected AI regulations and oversee the implementation of many of the organization’s AI ethics governance and education functions.

Several companies, such as BCG, Salesforce, IBM and Microsoft, already have this role with various titles – Head of Responsible AI, AI Ethics Global Leader, Global AI Ethicist, or Chief Responsible AI Officer. “[T]he call for artificial intelligence ethics specialists is growing louder as technology leaders publicly acknowledge that their products may be flawed and harmful to employment, privacy and human rights,” says a recent WSJ article.

At a very high level, companies need an AI ethics framework to ensure that AI-enabled solutions are developed in ways that mitigate the chances of harm to relevant stakeholders. More specifically, a CAIEO should lead the definition of broad AI ethics goals and then help the company understand how to meet these. They need to ensure that the AI technology being developed has suitable properties (fairness, robustness, explainability) and that developers have the right tools and training to easily embed these properties in what they produce. They must make sure that risks in AI deployment (both internally and to other companies) are appropriately mitigated.

IBM's trustworthy AI approach with clients
IBM's trustworthy AI approach with clients
Image: IBM

To achieve all this, the CAIEO role needs:

  • Multi-disciplinary knowledge: AI ethics issues cannot only be addressed through technical solutions and compliance with relevant policies, standards and laws. CAIEOs need multi-disciplinary knowledge and skills, including technical AI knowledge and perception, ethical considerations, social science and technology law familiarity, as well as business strategy know-how. They also need to facilitate the creation of tools and frameworks that enable product teams to develop AI responsibly, as well as to work with the whole company, other professional organizations and policy-makers to help shape laws, norms and standards to define and govern best practices at a global level.
  • Effective and inclusive governance: Companies can support and provide oversight for a CAIEO by creating an AI ethics board, led by the CAIEO with representatives from all the company’s divisions and with decision-making power, visibility and governance authority fully supported by the CEO and senior executives. The CAIEO can help each board member understand AI ethics solutions in their respective business units and make them part of operations. In large organizations, CAIEOs should implement top-down, centralized governance initiatives, such as corporate directives that state how the whole company should detect and mitigate AI bias when building or using an AI solution; and bottom-up initiatives, such as tools specific to a business unit or an AI solution.
  • Strategic differentiation and business value: Until now, much of the argument for CAIEOs has focused on risk reduction, but this role should also encourage companies to consider AI ethics as a source of value and a strategic differentiator rather than just a set of guardrails with which to comply. AI ethics principles and their implementation should be linked to the company’s values and business model so that all internal and external stakeholders can appreciate the value of developing, deploying and using AI responsibly. CAIEOs must understand the business value of investing in ethics and fairness, including costs for product development, implementation and business adoption.
  • Public communication and advocacy: CAIEOs also require communication skills to facilitate dialogue and trust between stakeholders within the company and externally. Helping people understand the issues and persuading them to change their actions requires strong communications skills and working across all aspects of the organization. A CAIEO must prepare content for targeted audiences and keep the dialogue and debates moving forward. AI ethics involves a deep understanding of regulations, governance and policy issues. Laws and regulations always lag behind technology and AI adoption, however, so it is essential to focus on values, norms and public perception.
  • Company-wide engagement: All this cannot be done by one person (or team) but instead requires a company-wide approach, where all business units contribute to achieving these goals. AI requirements need to be defined, technical tools built and educational materials produced. Clients need to be engaged and teams educated. According to the IEEE Ethically Aligned Design Guidelines for Autonomous and Intelligent Systems: “[C]ompanies need to create roles for senior-level marketers, ethicists or lawyers who can pragmatically implement an ethically aligned design, both in the technology and the social processes to support value-based system innovation.”

Data and AI are becoming core to most enterprises. While they reinvent their business model and strategic value in this new data-centric era, it is imperative that these organizations correctly and effectively identify and address AI ethics issues and risks. To do this – and to lead in creating both social and business value by building, deploying, and using AI in a responsible and trustworthy way – we advocate starting with the appointment of a CAIEO. This role will enable the creation of a company-wide approach to AI ethics.