Northeastern expert wants to move AI governance from principles to practice
On the London campus, Denise Garcia, a visiting professor from Boston, warns of the dangers of not properly regulating use of AI weaponry.
LONDON – In the last five years, AI has dramatically changed warfare in Ukraine, Israel and Iran by replacing humans with fully autonomous battle systems, accelerating decision-making and reshaping how targets are identified. Yet some experts worry that there remains a concerning lack of regulation for autonomous weapons and AI military systems.
At a recent public discussion on the London campus of Northeastern University highlighting the topic of “Breathing Oxygen into a Failing World Order: AI’s Global Diplomacy”, Denise Garcia, Professor of Political Science and International Affairs, shed light on the lack of any coordinated global regulation of AI.
Garcia, who is based on Northeastern’s Boston campus and is author of the book The AI Military Race, Common Good Governance in the Age of Artificial Intelligence, shared her thoughts during a moderated discussion.
Her comments have been edited for length and brevity.
How do you view AI today? What do you see as its main opportunities and risks?
The risk is frontier labs [leading companies experimenting with AI] that want to reach superintelligence. And I think that’s where we should put a stop point. There are military countries that are not yielding to the wishes of all the other countries that want regulation. I also see the environmental cost is mounting and I think we should be quite vigilant. I want to see a positive way out of this but I sometimes am very concerned – it’s a very grim moment in history, I’m sad to say.
What is the state of peace in the world as you see it?
We have now reached an astounding two figures here. One figure is over $2 trillion per year of [global] expenditure on the military. That for me is excessive. The other figure that we just came up with this year at the new Global Peace Index is almost $22 trillion of global economic impact of violence – that’s 10% of the global GDP. So that signals to us that too much is going to war, violence, all the conflicts that are happening. And the war in Iran will make that spike even higher for the next Global Peace Index. I can think of many more good uses of that for those trillions – more hospitals, more roads, public transportation. We see a diversion of humanity’s economic resources in a way that is astounding. We’re seeing the futility of war; war is no longer useful for solving anything. So it’s a cautionary tale here for us for the future of humanity.
What about AI? What role does AI play in the future of these trends?
Editor’s Picks
What I see happening in terms of AI impacting international legal architectures that we have, safeguarding humanity, the situation is not so good. With AI, countries think they can go to war quickly, win, and get out. But that’s not happening. We see in Iran, the quagmire that that became, we see Ukraine, Sudan. There is no easy fix because the ones that are being invaded can also defend against AI, with AI. International law is a social process and it’s human-centered. So AI complicates that equation. If something goes wrong, who created that? Who did that wrongdoing? Who do we attribute responsibility to? In the past 100 years, international law has been growing to benefit humanity and human security [and to] create more peace. Now, with AI, there’s a complication of the implementation of these legal architectures – lowering the threshold of war, putting civilians in the crossfires. The risks are getting larger and larger. We are entering an era of risks where AI could eliminate all of humanity.
When we look to the past, are there lessons we’ve learned from previous treaties that can help address AI?
Hopefully! The most successful international treaty that exists that I looked at for my book is the Montreal Protocol (1987) to repair the ozone layer. That protocol has 198 members, more than the [United Nations] membership. It was the largest repair mission that countries have undertaken together. The ozone layer is what enables life on Earth, it enables us to be here and not burnt by the sun. The Montreal Protocol stands the test of time, it evolves with scientific discoveries of new harms that are posed to the ozone layer. Every meeting of the member states every year is very joyous because they know this is the most successful endeavour. They think about how they can adjust the treaty. So that would be a good lesson for AI governance.
What do you see as the road ahead within these diplomatic discussions around AI?
I want to be optimistic because we are Northeastern University. At Northeastern, we are science-based optimists. I think this year and next year it’s going to be quite pivotal actually. We are living in a transformation right now. We had a few resolutions at the UN General Assembly that signalled to me that there’s a growing mounting consensus that we need a framework convention that will have several treaties under it, because AI is so vast. We had the first global dialogue in Geneva on AI [in July] where Northeastern was present. We’re going to be present as Northeastern again in the second dialogue next summer. There is a lot of evil, a lot of concern. But there is that much good going on as well and that energises me a lot.
