Read the following passage and mark the letter A, B, C, or D to indicate the correct answer to each of the questions from 34 to 38.

When it comes to discussing AI and the future of work we exhale in relief when someone says machines will have a hard time replacing us. This statement is backed by the argument that most AI systems are ‘narrow’. AI systems only do one thing but do it really well, for example, predicting what you want to watch on Netflix. If you ask the same system to make you a cup of coffee or drive your car, you’re likely to be disappointed.

But what happens when the AI system can perform the most significant tasks that a job entails? One interesting example we found came from the fashion industry, where a company had implemented two AI systems to produce novel designs. Together, they did all work and the human’s role was only to surveil the work.

Research published earlier this year coined the term ‘Shadow Learning’. As explained in the Harvard Business Review article titled ‘Learning to work with intelligent machines’, the researcher studied the challenges new surgeons faced when learning robotic surgery skills. Previously, they learned how to perform surgery by working alongside expert surgeons, but now they’re forced to watch over the surgeon’s shoulder as, thanks to robotics, individuals can handle entire surgeries with one pair of hands. The term, however, does not mean that you learn by shadowing someone. The phrase refers to students who gained experience with robotic tools by taking it upon themselves to acquire new skills outside of the curriculum.

If we look beyond the current discussion of AI and the future of work – which usually revolves around the number of jobs that will be impacted by AI – we can focus on how to create inspiring new ways to work with machines.

(Source: https://www.ericsson.com/)

According to paragraph 3, what is the problem for surgical apprentices in the time of AI universality?