ABSTRACT
Artificial intelligence is increasingly able to do tasks previously performed by humans. This has significant implications for both e-cheating and assessment security. Students can increasingly ‘offload’ tasks like paraphrasing or algebra solving to computers, which makes assessing students’ capabilities difficult. Conversely, computers are increasingly used to undertake assessment security tasks, like artificial intelligence proctored examinations. Five principles are proposed for evaluating the use of artificial intelligence in an assessment security context, including explainability and transparency, accuracy, fairness, accountability and empathy. A key question in this chapter asks which tasks should remain human if artificial intelligence is able to take on complex content production and assessment security tasks.
