Great Powers Are Failing Us on AI. It's Time for Small States To Step In | Opinion

A recent open letter calling for a six month moratorium on training artificial intelligence (AI) systems more powerful than GPT-4 has provoked a controversy. Among the letter's many illustrious signatories is Elon Musk, who days later announced his own AI project. His name has ignited accusations of ulterior motives. This squabble is not only needless and unfortunate—it is also a stark demonstration of our limitations as human beings, and evidence of the urgent need for raising awareness of AI and creating thoroughgoing regulations and enforcement mechanisms around it.

In my lives as a scientist, diplomat, businessman, and politician, I have been a close observer of epochal events—from the collapse of the Soviet Union and the emancipation of suppressed nations, to the spread of technology, and the empowerment of marginalized individuals. Each achievement was undermined by a very human failure to grapple sincerely with its ramifications. Let's be clear: AI is a monumental accomplishment. The technology we now possess is perhaps as significant as nuclear power. It can rapidly uplift humanity, but, in the absence of sustained human oversight, it can also plunge us into serfdom. What we do now, in this fleeting moment, will determine our fate.

Let us first dispense with the fearmongering. AI has the potential to revolutionize virtually every aspect of human life—from manufacturing, education, and health care, to finance, and transport. All of this will not only increase productivity and improve our lives, but also liberate the multitudes currently trapped in what the author George Orwell called a "routine which makes thought impossible."

We should not be afraid of AI. We should be seriously worried about the abuse and misuse of it. The objective of automation should be to enhance the quality of human life. The human beings whose jobs are vulnerable to this technological advance cannot be treated as a minor issue. Nor can we afford to overlook the implications for privacy as vast amounts of data are fed for analysis to AI machines controlled by big corporations and governments.

AI has made it theoretically possible for rogue states to develop new ways and means of threatening larger powers. AI-run machines will almost certainly be engaged to manipulate public opinion and stage cyber attacks on critical infrastructure.

Given these and numerous other risks, it is imperative that the development of AI is regulated and monitored by a transparent international body. This would require global cooperation and consensus-building. The current international climate is not conducive to such an endeavor. Against this backdrop discord, any domestic regulations that fetter the further development of AI in the U.S. or Europe alone, while leaving powers such as China and Russia to pursue their aims freely, will be tantamount to self-harm.

Domestic regulatory activity around AI must be focused primarily on averting social unrest and preserving human dignity. The objective of the technology must be to promote human well-being, and not just to maximize efficiency or profits. AI must not be allowed to be used to perpetuate or exacerbate existing inequalities.

The emergence of new technologies always leads to significant changes in military strategies—in planning and the organization of armed forces. And it is impossible to prevent the use of artificial intelligence for military purposes. We are entering a new global technology race that will lead to the introduction of the most modern innovations in the military sphere in the foreseeable future. All major world powers will seek to exploit this innovation, and those who lag behind will expose vulnerabilities that will be virtually impossible to shield with conventional weapons. My late friend Stephen Hawking warned about a future in which AI would enslave humanity. I fear a future in which AI leads great powers inadvertently into annihilating humanity.

Militarily, AI is not a challenge that any single nation, much less a tech firm, can solve on its own. It should provoke us into thinking of ourselves as human beings first and devising rules that protect our interests as a species. The energies of governments must be directed urgently into creating something akin to a non-proliferation regime for AI to ensure that non-state actors do not acquire capabilities that could be put to dangerous uses.

Given the stakes, the failure of great powers to regulate AI is an indication of their blindness to their own impending obsolescence. Can they set aside their differences and convene a global convention on regulating AI? Since the answer isn't certain, and since the world doesn't have the luxury of waiting, the burden of mobilizing a global response commensurate to the challenge of AI may have to be borne by small powers.

I have long been a proponent of a club of smart and responsible small states—an idea that hasn't impressed multilateral bodies comfortable with the status quo. But the status quo is untenable. Since the advent of the atomic bomb, balance of power in international relations has rested on the dominance of nuclear powers, which form the rules of the game and set out the red lines for all others. AI has created an important opening for small states to play a constructive role, and this is the moment for such states to unite and assume the leading role in convening a global conference, canvassing and mediating between quarrelling big powers, and brokering an international treaty on AI.

The only solution is global dialogue, facilitated and convened by responsible players to frame and systematize rules around AI. What needs to pause is not the growth of technology but the feuding among nations, from Europe to Asia, that is undermining our ability to harness human achievements to benefit humanity.

Dr. Armen Sarkissian served as the fifth prime minister and fourth president of the Republic of Armenia. His next book, Small States Club: How Small Smart States Can Save the World, will publish this year.

The views expressed in this article are the writer's own.