If they aren't already familiar with AI risk, then they probably won't read Nick Bostrom's Superintelligence (2014) or AI as a Positive and Negative Factor in Global Risk (2008). For people whom a single article is more appropriate, keeping in mind the Lessons Learned from Talking to Academics about AI Safety, what is the best resource for introducing someone to AI safety?
What is their level of familiarity with machine learning and/or computer science?
High on computer science, very low on ML.