I glitters on the global stage. The meteoric rise of AI-chip manufacturing company—Nvidia—dazzles world leaders and financial wizards. It enthrals even mum-and-dad retail investors, whose eyes turn into dollar signs as they spot a once-in-a-lifetime opportunity to make a killing in the stock market.
Virtually unknown 15 years ago, Nvidia, with a $5 trillion market capitalisation, is now the world’s most valuable company. It is not comparable, but to get a sense of the enormity of this figure, Nvidia’s valuation is higher than the GDP of all countries in the world, except two—the US and China. A $10,000 investment in Nvidia five years ago is currently worth $137,000, an astounding 1,270 per cent return. Three-fourths of its employees are dollar millionaires.
This money-mania drowns critics who warn about AI’s “stealth-and-steal” persona. AI’s strengths are well-documented, but regulators fret over the excessive concentration of wealth and power, its impact on politics, academia and society. AI accelerates job losses, tramples truth, steals creativity and distorts culture. An all-too human way of life threatens to fade away.
Illustration: Deni Lal
The arguments that AI cannot match human creativity, decision-making and empathy—the hallmarks of our species—lie like wrecks in a war zone. Haiku, the 17-syllable Japanese poem, is breathtaking in its brevity and brilliance. Basho remains the haiku maestro. A decade ago, AI’s haiku renditions were laughable. Now, they stun. It was said, AI is no Bach. Perhaps, but it can imitate so well that music lovers in a Los Angeles salon could not distinguish AI and Bach compositions. In our arrogance, we failed to recognise a fundamental revelation: AI improves constantly and exponentially. It corrects its mistakes faster than humans. Hubris meets Nemesis.
The current sensational phenomenon is AGI (Artificial General Intelligence). Warring nations use the euphemism “collateral damage” to disguise the brutality of civilian casualties. AGI disguises its real meaning: the turning point when AI matches human intelligence. Then comes AI’s Superintelligence, surpassing human intelligence. This will be the pinnacle of human achievement. Or, downfall.
Mortal accomplishments leave in their wake winners and losers. The rich get richer, Nvidia reaches $10 trillion market cap. But losses are inevitable. Bubbles burst, companies collapse, bankruptcies spiral, exuberance crashes. Humanity’s capacity for self-harm is explained by the ‘Prisoner’s Dilemma’ conceived by mathematicians Melvin Dresher and Merrill Flood in 1950. If two arrested criminals cooperate and remain silent, the outcome is good. They get short sentences. If both confess, they get moderate sentences. But if one betrays, while the other remains silent, the betrayer is freed, a fantastic outcome for him, but his partner gets a long sentence. The obvious choice for individuals is to pursue self-interest, by betraying, not cooperating.
The dilemma is that the high returns of self-interest invariably trumps the moderate returns of collective good—not just for criminals, but in economics, psychology, geopolitics or environment negotiations. Self-interest is particularly pronounced in AI. Cooperation entails regulation, investing in safety, sharing progress. But this requires unprecedented and unimaginable global trust and enforcement. So, AI safety sputters while Big Tech’s investment in AI skyrockets.
Experts say, once one company or country reaches superintelligence, the race ends. Competitors cannot catch up because the winner exponentially outpaces them. The victor amasses colossal financial, political, even existential benefits. Others are left with little or nothing. It is catastrophic for losers, but euphoric for winners, even beneficial for humanity if superintelligence is humane. ‘The Winner Takes It All’ is not just an Abba song. It could be a way of life.
Pratap is an author and journalist.