A、It needs instructions throughout the process. B、It does poorly on frequent, high-volume tasks. C、It has to rely on huge amount

Here is my baby niece Sarah. Her mum is a doctor and her dad is a lawyer. By the time Sarah goes to college, the jobs her parents do are going to look dramatically different.
    In 2013, researchers at Oxford University did a study on the future of work. [16] They concluded that almost one in every two jobs has a high risk of being automated by machines. Machine learning is the technology that’ s responsible for most of this disruption. It’ s the most powerful branch of artificial intelligence. It allows machines to learn from data and copy some of the things that humans can do.
    My company, Kaggle, operates on the cutting edge of machine learning. We bring together hundreds of thousands of experts to solve important problems for industry and academia. This gives us an unique perspective on what machines can do, what they can’t do and what jobs they might automate or threaten.
    Machine learning started making its way into industry in the early 90s. It started with relatively simple tasks. It started with things like assessing credit risk from loan applications, sorting the mail by reading handwritten zip codes. Over the past few years, we have made dramatic breakthroughs. Machine learning is now capable of far, far more complex tasks.
    In 2012, Kaggle challenged its community to build a program that could grade high school essays. [17] The winning programs were able to match the grades given by human teachers. Now given the right data, machines are going to outperform humans at tasks like this. A teacher might read 10,000 essays over a 40-year career. A machine can read millions of essays within minutes. We have no chance of competing against machines on frequent, high-volume tasks, but there are things we can do that machines cannot. Where machines have made very little progress is in tackling novel situations. Machines can’ t handle things they haven’ t seen many times before. [18] The fundamental limitation of machine learning is that it needs to learn from large volumes of past data. But humans don’ t. We have the ability to connect seemingly different threads to solve problems we’ ve never seen before.
Questions 16 to 18 are based on the recording you have just heard.
16. What did the researchers at Oxford University conclude?
17. What do we learn about Kaggle company’ s winning programs?
18. What is the fundamental limitation on machine learning?