Like code, poems struggle to capture meaning artfully via methods and metaphors within a specific meter and/or structure. Whereas code is meant to be run, poems are meant to be read. But unlike code, poems usually deal explicitly with the emotional and often ethical. Ethics often struggles to become experiential. At the same time, given that the world is increasingly experiencing services provided by artificial intelligence (AI), AI often struggles to be related to the ethical. It is becoming more and more apparent, artificial intelligence threatens to irreversibly change the world. Given the power, acceleration, and susceptibility (to bias, adversarial attacks, and much more) of artificial intelligence, discussion  on the ethics of its use seems lacking.

Given this opportunity, I propose a poetry competition to bring more attention and thought to the ethics of artificial intelligence.

Inspired by “Do Artifacts Have Politics”,  contestants can answer questions like “Is AI inherently political?”, followed by examples of how AI can result in negative side effects, reward hacking, and blind spots in application to the real-world (Winner 1980). Like in politics, any one AI case can represent the constituent AI field, and any artifact is intentionally designed. Participants can describe how lapses in integrity can lead to dystopian environments, satirize current problems caused by AI, or otherwise make broad sweeping claims on whether AI is arguably responsible for humanity's future.

Similarly, using ethical lenses as a brainstorming tool for capturing the different ethical dimensions of artificial intelligence, scientists can use first principles like universalizability, consequence, character, and justice to describe emergent models to judge AI. Similarly, outcomes can be measured by frameworks like the Greater Good canvas, which questions design, expected value, and behavior in an ethical risk assessment. These benchmarks or goals are directly applicable to ensure oversight, safe exploration, and robustness to distributional shift.

Finally, the age-old question of whether AI is truly AI remains. Under the modern conception of Artificial Moral Agents, like described in “Artificial virtue: The machine question and perceptions of moral character in artificial moral agents”, we can return to “classical” philosophy, like the  and paternal utilitarianism. For example, the Turing Test provides an enduring benchmark and experiment to weigh the human-ness of AI. An array of speculative fiction on the subject, especially Asimov’s Three Laws of Robotics and “I-robot : Poetry”, serve to inspire.