On Monday, US President Joe Biden released a far-reaching and impressive executive order on artificial intelligence (AI) – catapulting the US to the forefront of conversations about AI regulation.
This puts the USA ahead of other states within the race for dominance over AI. Europe previously led the way in which with its AI law, which was passed by the European Parliament in June 2023 but is not going to take full effect until 2025.
The presidential executive order is a plethora of initiatives to control AI – a few of them good, others seem slightly half-baked. The aim is to combat harms starting from immediate harms comparable to AI-generated deepfakes, to medium-term harms comparable to job losses, to longer-term harms comparable to the much-debated existential threat that AI can pose to humans.
Biden's ambitious plan
The U.S. Congress has been slow to pass comprehensive regulations for giant tech corporations. This executive order is probably going as much an attempt to bypass an often-gridlocked Congress because it is to spur motion. For example, the order calls on Congress to pass bipartisan privacy laws.
Bipartisan support in the present climate? Good luck with that, Mr. President.
The executive order is reportedly expected to be implemented inside the subsequent three months to a yr. It includes eight areas:
- Security standards for AI
- data protection
- Justice and Civil Rights
- Consumer rights
- Jobs
- Innovation and competition
- international leadership
- AI governance.
On the one hand, the regulation covers many concerns raised by scientists and the general public. For example, one in every of their policies is to issue official guidance on how one can watermark AI-generated content to cut back the danger of deepfakes.
Additionally, corporations developing AI models must exhibit their security before they may be rolled out for wider use. President Biden said:
That means corporations must inform the federal government concerning the large-scale AI systems they’re developing and supply rigorous independent testing results to prove they don’t pose a national security risk to the American people.
The potentially catastrophic use of AI in warfare
At the identical time, the order doesn’t address quite a few pressing issues. For example, it doesn't directly address how one can take care of killer AI robots, a vexed topic discussed on the United Nations General Assembly over the past two weeks.
This concern shouldn’t be ignored. The Pentagon is developing swarms of low-cost autonomous drones as a part of its recently announced Replicator program. Similarly, Ukraine has developed homegrown AI-powered attack drones that may discover and attack Russian forces without human intervention.
Could we find yourself in a world where machines determine who lives or dies? The executive order only requires that the military use AI ethically, but doesn’t specify what which means.
And what about protecting elections from AI-powered weapons of mass persuasion? Several media outlets have reported that the recent elections in Slovakia could have been influenced by deepfakes. Many experts, including myself, are also concerned concerning the misuse of AI within the upcoming US presidential election.
Without strong controls, we risk living in a time where nothing we see or hear online may be trusted. If that sounds exaggerated, consider that the US Republican Party has already released a campaign ad that appears to have been generated entirely by AI.
Missed opportunities
Many of the initiatives contained within the Executive Order could and needs to be replicated elsewhere, including Australia. We must also, as required by the regulation, provide guidance to landlords, government programs, and government contractors on how one can be certain that AI algorithms will not be used to discriminate against individuals.
We must also, as required by the order, address algorithmic discrimination within the criminal justice system, where AI is increasingly getting used in high-risk situations, including for sentencing, probation and parole, pretrial release and detention, risk assessments, surveillance, and predictive policing, to call just a few.
AI can be getting used controversially for such applications in Australia, for instance within the Suspect Targeting Management Plan for monitoring young people in New South Wales.
Perhaps probably the most controversial aspect of the manager order is the one which addresses the potential harms of probably the most powerful so-called “frontier” AI models. Some experts imagine these models – developed by corporations like Open AI, Google and Anthropic – pose an existential threat to humanity.
Others, including me, imagine such concerns are overblown and will distract from more immediate harms like misinformation and inequality which can be already harming society.
Biden's order invokes extraordinary war powers (specifically the Defense Production Act of 1950, established throughout the Korean War) to require corporations to notify the federal government after they form such frontier models. They must also share the outcomes of “red team” security tests, where internal hackers use attacks to look at software for errors and vulnerabilities.
I’d say that it’s going to be difficult and even perhaps inconceivable to watch the event of boundary models. The above guidelines is not going to prevent corporations from developing such models abroad, where the ability of the US government is restricted. The open source community can even develop them in a distributed manner – making the tech world “borderless.”
The impact of the manager order will likely have the best impact on the federal government itself and its approach to AI, slightly than on corporations.
Still, it's a welcome move. British Prime Minister Rishi Sunak's AI security summit, which can happen over the subsequent two days, appears to be something of a diplomatic talk-fest by comparison.
It makes you jealous of the president's power to get things done.
