Never in the history of humanity have we allowed a machine to autonomously decide who should live and who should die, in a fraction of a second, without real-time supervision. We are going to cross that bridge any time now, and it will not happen in a distant theatre of military operations; it will happen in that most mundane aspect of our lives, everyday transportation. Before we allow our cars to make ethical decisions, we need to have a global conversation to express our preferences to the companies that will design moral algorithms, and to the policymakers that will regulate them.

– Awad et al. [2018, p. 63]

Artificial intelligence is a transformational set of ideas, algorithms, and tools. AI systems are now increasingly deployed at scale in the real world [Littman et al., 2021; Zhang et al., 2022a]. They have significant impact across almost all forms of human activity, including the economic, social, psychological, healthcare, legal, political, government, scientific, technological, manufacturing, military, media, educational, artistic, transportation, agricultural, environmental, and philosophical spheres. Those impacts can be beneficial but they may also be harmful. Ethical and, possibly, regulatory concerns, as raised by Awad et al. [2018], apply to all the spheres of AI application, not just to self-driving cars.

Autonomous agents perceive, decide, and act on their own. They, along with semi-autonomous agents, represent a radical, qualitative change in technology and in our image of technology. Such agents can take unanticipated actions beyond human control. As with any disruptive technology, there may be substantial beneficial and harmful consequences – many that are difficult to evaluate and many that humans simply cannot, or will not, foresee.

Consider social media platforms, which rely on AI algorithms, such as deep learning and probabilistic models, trained on huge datasets generated by users. These platforms allow people to connect, communicate, and form social groups across the planet in healthy ways. However, the platforms typically optimize a user’s feed to maximize engagement, thereby increasing advertising revenue. Maximizing engagement often leads to adversarial behavior and polarization. The platforms can also be manipulated to drive divisive political debates adversely affecting democratic elections, and produce other harmful outcomes such as deep fakes. They can also be very invasive of users’ privacy. Automated decision systems, possibly biased, are used to qualify people for loans, mortgages, and insurance policies, and even to screen potential employees.

People expect to have the right to receive fair and equitable treatment, to appeal decisions, to ask for accountability and trustworthiness, and to expect privacy. Is it possible to ensure that those rights are indeed available?