Robots have once again proven their ability to outperform humans. This time, it’s in a way that reminds us of their potential to replace human workers.
Two artificial intelligence (AI) algorithms developed by Microsoft and Chinese marketplace Alibaba defeated humans at a global reading test called the Stanford Question Answering Database. The exam was developed to track the progress of AI’s reading abilities. It’s similar to the reading portion of standardized tests in that it asks AI to read a paragraph and answer questions about it. In this case, robots read 500 Wikipedia articles and answered more than 100,000 questions.
Alibaba’s research unit, the Institute of Data Science of Technologies, said its deep-learning model achieved a score of 82.44, narrowly beating out the human benchmark of 82.304. It marked the first time in history that a machine outscored a human in this type of reading test. A day later, Microsoft’s Asia research team joined Alibaba, beating the humans with a record score of 82.650.
Divisions from other major companies and researchers at universities had previously failed to surpass the human score. Among those that attempted are Facebook, Samsung, IBM, Google, Tel-Aviv University, and Carnegie Mellon University. Many contestants completed the test multiple times, tweaking their algorithms after each attempt. Both Alibaba and Microsoft scored below 80 points on previous runs.
Unlike AI built to defeat humans in Go or Starcraft, these intelligent machines are taught skills required in human occupations. There’s a real fear that a significant number of workers will soon lose their jobs to intelligent machines. In fact, 40 percent of jobs in the U.S. are at high risk of being replaced in the next 12 years, consulting firm PwC predicts.
Alibaba has already used the underlying technology of its reading AI to build a chatbot for its Black Friday-style shopping festival that could respond to customer inquiries. The company said it would expand its application in the near future and will share its model-building methodology with the community.
“That means objective questions such as ‘what causes rain’ can now be answered with high accuracy by machines,” Si Luo We, Alibaba’s chief scientist of natural language processing, told ZDNet. “We believe the technology underneath can be gradually applied to numerous applications such as customer service, museum tutorials, and online responses to medical inquiries from patients, decreasing the need for human input in an unprecedented way.”
Robots are already driving humans out of jobs at alarming rates. To prevent unemployment rates from increasing any further, former CEO of Microsoft Bill Gates proposed that robots should be taxed. However, the company he founded believes robots and humans can work together.
“These kinds of tests are certainly useful benchmarks for how far along the AI journey we may be,” said Microsoft spokesperson Andrew Pickup. “However, the real benefit of AI is when it is used in harmony with humans.”