Future Intelligence exclusive from the Benevolent Artificial General Intelligence Conference in Panama
Some of the world’s leading ethicists are calling for a supranational body to control AGI development in similar ways to those used to oversee, nuclear weapons and genetics.
The push for some form of control of Artificial General Intelligence, a form of the technology likely to be able to approach awareness, which started at the Benevolent Artificial General Intelligence Conference in Panama in March, has come amid growing concerns about the uncertain impact of the technology.
Chief among those concerns in the short-term are job losses to AI but much deeper fears about a loss of control to a super-intelligent system, inability to understand how it works and a loss of human confidence in the face of such technology have also emerged according to the philosopher and futurist Jerome Glenn, the Director of the Millennium Project who was one of the speakers at the event.
“One of my concerns is that people are not aware of how soon AI could evolve beyond what they think today, so they’re not taking seriously what it might look like and what sort of rules we might need. There have been meetings of UNESCO and there’s been OECD. The White House has got some rules. Chinas got some rules. The EU’s got some rules, but they’re mostly focused on the narrow intelligence we have today. They’re not anticipating the future artificial general intelligence that we can have tomorrow,” said Glenn, a former associate of the renowned thinker Alvin Toffler, and one of the architects of the Strategic Arms Limitation Talks which limited nuclear weapons in 1979.
Potentially dangerous AI too close for comfort
Speaking in an exclusive interview with the Future Intelligence’s, ‘PassW0rd‘ radio programme ‘In AI We Trust’, Glenn went on to say why there was a need for a similar treaty to the SALT debate to govern AI and specifically Artificial General Intelligence rather than the narrow general intelligence we think of as AI that we have become aware of in the last two years.
“Several years ago, we thought, artificial general intelligence is ten, twenty, thirty, forty, fifty years from now. Now people aren’t saying that it’s much sooner, so to get national rules and UN rules it would be a miracle if we got them done in a couple of years, but that’s what we’re going to have to do.
“If AGI could be released in, say, three to five years, and if we don’t have something in place that helps manage that, how it gets done and released, then unregulated AGI by many different countries it seems to me it’s not likely that will turn out OK and that could be a mess for future civilizations. So, to prevent that, we got to have serious conversations about future rules, about systems that don’t exist yet.”
Glenn, whose Millennium Project is backed by Vint Cerf, one of the architects of the internet, is making representations to the UN and a number of other extra-national organisations in a bid to bring about a supranational body to control AGI development.
An attempt beginning to receive a great deal of support since concerns about AI technology were first mooted by Risto Siilasmaa, the Finnish former chief executive of Nokia and the cybersecurity company WithSecure, who went on a self-funded world tour to alert national leaders of the danger from the technology.
A constitution for AGI and supranational body
Joining Glenn in Panama was the lawyer, and entrepreneur, Anneloes Smitsman. Smitsman, whose latest venture is an ethical computer game called Elowyn, said there was an urgent need to bring together the various competing AI regulation initiatives in the US, the EU, China and the UK all under one umbrella to address the issue of AI development. To that end, Smitsman is developing a framework with the aim of creating a global constitution for AGI development.
“That’s exactly why this participatory framework for guiding the creation of an AGI constitution is about. It’s to say let’s unite, let’s put it all together. Let’s really look here. What is complementary?
“Also, to understand that the approach towards narrow AI may not work for AGI and I don’t think that’s sufficiently on the agenda or even in people’s minds so.
“I think one of the danger points could be to think that we’ve got this figured out now with narrow AI, and we have regulations in place and there are agreements here and then the baby AGI emerges, and it’s a completely different ball game.
“So, we are proposing to stretch our thinking around this and to start to come towards, well agreed, formulated and implementable common good governance principles,” said Smitsman who has set up a steering committee to present the framework to organisations like the UN, the OECD and the EU.
For more on AI regulation and a supranational body to control AGI, see:
AI experts call for regulator – Future Intelligence
Techlash: the end of the internet dream? – Future Intelligence
Top tech lawyer calls for Roman Slave Law and agency to control AI use – Future Intelligence
