Law and Order on the AI Frontier
Two professors examine advances in automation, robotics and artificial intelligence, as well as the related uncharted regulatory and ethical territory.
- Joanna Bryson, Associate Professor at the University of Bath and Researcher at Princeton University
- Aimee van Wynsberghe, Assistant Professor of Ethics and Technology, Delft University of Technology
- Moderator: Rob Cox, Founder and Global Editor, Breakingviews
International, regional and national agencies, including the United Nations and the European Union, are working on various AI initiatives, and much remains to be done. AI must conform to existing laws and regulations concerning liability and privacy; prosecutors and regulators need to be trained; gaps in the law must be identified and addressed by policymakers; and new ethics debates must take place.
What is AI?
Bryson believes artificial intelligence boils down to the ability “to do the right thing at the right time.” Thus, AI can be identified as the transition from computation to action. “AI is not just maths and computation,” she said. “It takes up time, space and energy.”
AI can also come in different forms. It can be defined as intelligence embodied in a physical entity that shares our environment. Or it might be an avatar on a screen, with which people interact, but not physically. It could also be the software for an online chatbot.
Human beings “are good at learning, compressing knowledge, and passing it on,” Bryson said. From this perspective, AI consists of “all of our intelligence that we have uploaded.”
Legislating and regulating AI
If AI is embedded in a physical object, both the object and the software need to be regulated, van Wynsberghe explained. Moreover, legislation is required to regulate data quality, safety and privacy.
Bryson highlighted three main areas of concern:
- Faulty products – for example, an algorithm designed for use in the criminal justice system, which claimed to be able to predict recidivism, turned out to be both inaccurate and racially biased.
- Economic fairness and taxes – for example, people are handing over their data in exchange for services, meaning a lot of untaxed bartering is going on.
- Liability – for example, what happens when a driverless car causes injury or damage? Van Wynsberghe raised issues related to human interaction, such as the right to know if you are speaking with a human being when you contact a call center.
“We need knowledgeable prosecutors,” said Bryson. “We should mandate that AI conform to present law.” Like the pharmaceutical industry, AI and robotics should be regulated “for the entire lifecycle of the product,” argued van Wynsberghe. Bryson compared eventual AI law to the construction industry, saying that 100 years ago, anyone with enough money could put up whatever they wanted. Today we have zoning laws and building codes.
Explore the Summit
Speaker opinions may not necessarily be those of Natixis Investment Managers. Not all speakers are employed by Natixis Investment Managers, but may receive compensation for their services. Content should not be considered a solicitation to buy or an offer to sell any product or service to any person in any jurisdiction where such activity would be unlawful.
The Natixis Investment Managers Summit was hosted by Natixis Investment Managers. Natixis Investment Managers includes all of the investment management and distribution entities affiliated with Natixis Distribution, L.P. and Natixis Investment Managers S.A.