The World Economic Forum wants to create an "ethics switch" to prevent artificial general intelligence from being harmful or unethical.
Dan Patterson, a Senior Producer for CBS News and CNET, interviewed Murat Sönmez, director of the World Economic Forum, about the future of artificial general intelligence (AGI). The following is an edited transcript of the interview.
Dan Patterson: AGI: When?
Murat Sönmez: It's hard to predict. A lot of initiatives like OpenAI are working on that with a billion dollar investment from Microsoft. The question is, until then, what did we do? Because the threats--these algorithms--are not written by a human being, and it's derived from the data, we need to make sure that data is diverse enough, and it doesn't amplify our biases already in this society, and it's from real sources, not fake data.
Let's say that we have it, but there's also a real chance that these algorithms will do something that's harmful or unethical because we have no control over it. The question is, how do we prevent it from harming the society?
SEE: An IT pro's guide to robotic process automation (free PDF) (TechRepublic)
We looked back in the history of technology, in the industrial revolutions, and said, 'Has this problem been addressed before?' And it was, when we switched from steam to electricity. US factories adopted electricity in the late 19th, early 20th century, and it caused all kinds of fires because the voltage was not regulated. The insurance companies in the United States got fed up paying for the damages, and they discovered the circuit breaker, which was invented by [Thomas] Edison, but there was no standard, so they created Underwriters Laboratories, a captive body, and they told the factory owners, 'If you don't use UL-certified circuit breakers, we're not going to insure you.' That's how we have safe electricity in the world. That's how this is being filmed and broadcast.
We said, 'Can we do the same thing?' We have a concept of an ethic switch where countries define their ethics rules. We download these rules to the smart devices. When they're asked to do something that's harmful, the switch will say no.
The manufacturers love it because it gives them a safe way to scale technology. Governments love it because it gives them a way to minimize harm. Insurance companies love it because they can write policies to force the usage of this ethics switch. Until AGI happens, we would like to create the circuit breaker for harmful actions.
Watch more interviews with Dan Patterson and Murat Sönmez
- Top 5 things to know about the state of artificial intelligence (TechRepublic)
- Telemedicine, AI, and deep learning are revolutionizing healthcare (TechRepublic download)
- Artificial intelligence ethics policy (TechRepublic Premium)
- AI explained: Everything you need to know about our robot overlords (CBS News)
- OpenAI's 'dangerous' AI text generator is out: People find GPT-2's words 'convincing' (ZDNet)
- 2020: The year of seeing clearly on AI and machine learning (ZDNet)
- 6 ways to delete yourself from the internet (CNET)
- Artificial Intelligence: More must-read coverage (TechRepublic on Flipboard)