The robot revolution is coming, and people like Elon Musk and prominent scientists are concerned about what this could portend for humanity. They have suggested that we carefully think through what it would mean to live with artificial sentiences that could easily outstrip our own intelligences, and whether it’s really worth bringing them into the world to make our chat apps a little better.
As part of a lengthy June 28 op-ed for Slate on the future of AI, Microsoft CEO Satya Nadella outlined six imperatives for artificial intelligence, and four things that humans must impart into AI systems they build. (Nadella is also working on a book that will touch on this subject.)
A quick summary of Nadella’s beliefs—10 commandments, if you will—regarding how AIs and humans must act in our shared future:
1. “A.I. must be designed to assist humanity”
Nadella’s op-ed references science-fiction writer Isaac Asimov’s often-cited three laws of robotics, which essentially state that robots should not harm humans. Nadella takes this slightly further in suggesting their programs should be designed so that they’re actively helping people. They should be, he writes, “collaborative robots, or co-bots.”
2. “A.I. must be transparent”
And humans must understand it. “The tech will know things about humans, but the humans must know about the machines,” he argues.
3. “A.I. must maximize efficiencies without destroying the dignity of people”
Technology, and AI, should promote diversity and the interconnectedness of humanity, rather than drive it apart—which humans themselves have historically been unfortunately excellent at doing.
4. “A.I. must be designed for intelligent privacy—sophisticated protections that secure personal and group information in ways that earn trust”
In other words, AI shouldn’t leave our privacies exposed.
5. “A.I. must have algorithmic accountability so that humans can undo unintended harm”
Basically, AI needs a kill switch in case it does something nefarious that its programmers didn’t expect it to do.
6. “A.I. must guard against bias, ensuring proper, and representative research so that the wrong heuristics cannot be used to discriminate.”
AI should represent the will of everyone, not just a specific group.
7. Proceed with empathy
Moving on to Nadella’s also four “musts” for humans, and the skills that we ought to pass along to future generations of AI architects, we start with empathy. Nadella says that to create robots that truly understand us, we’ll first need to ensure that we understand everyone else’s “thoughts and feelings.”
8. Invest in education
We will need to develop and finance training models to be able to facilitate the benefits that AI may bring to humanity. As Nadella points out, “The power loom was invented in 1810 but took 35 years to transform the clothing industry because there were not sufficient trained mechanics to meet demand.”
9. Go forth creatively
Humans will need to continue to create, as that’s one of the things that makes us human, but machines will be able to “augment” our creative output—something a lot of AI researchers are looking into.
10. Reserve the right of judgment and maintain accountability for it
“We may be willing to accept a computer-generated diagnosis or legal decision,” Nadella’s writes, “but we will still expect a human to be ultimately accountable for the outcomes.”