The seven ugly sins of AI: Left unregulated, they could cause hell

We are believers in innovation and think the long-term outcome will be positive. But faith cannot be blindness. This is not a crisis of faith; it is a demand for sight. At today’s valuations, ignoring risks would be the greater sin. To help track those risks, here are the ‘seven deadly sins’ of AI. Each is real, present and must be reckoned with.

Pride—Hubris of the machine age: We have given the machine a voice. It speaks our language, and some believe it is beginning to form thoughts. The true sin is not this private belief, but the public performance of absolute certainty that leadership now seems to require.

CEOs who once urged caution now preach revolution. As budgets for roads and hospitals are trimmed and non-AI private spending is pruned, much is being sacrificed at the altar of AI. Pride is the armour a leader must wear to justify all this.

The danger is not failure itself, but the spectacle of it. Everyone knows innovation projects can fail. Autonomous driving will not be perfect for years. Chatbots will cause harm. Each mistake will be broadcast, litigated and used as a weapon by rivals.

But the market does not reward nuance; it demands conviction. Those putting in billions must project a faith with no room for doubt. So every stumble will be called a fall and the leaders of this age will walk with a target pinned to their backs.

Greed—A mania that follows the money: AI is now the largest capital magnet in technology, and greed seems to have been given a new name: innovation. Startups are no longer born of vision, but built for acquisition. Their goal is not to create the next Facebook, but be snapped up for big money, even as insider rounds inflate valuations for the same investors to mark up their own assets on paper.

The giants make rushed acquisitions and sign poorly vetted data deals, spending billions in shareholder money to stay ‘in the race.’ Governance is blindfolded; boards approve moonshots with no maps to profit. But the bigger sin is the new accounting. A tech giant invests billions in a partner, which then pays those billions back for cloud services. Investment is booked as revenue, a closed loop of capital, like a serpent eating its own tail.

The 2008 crisis reborn in the cloud? Could be. Countless can sense it. If fortunes turn, the sages of doom will be those who happen to speak the loudest the day before reversals begin.

Lust—Extremism of the crowded stage: To be heard now requires a shout. So the lust is no longer for participation, but for relevance. And in a crowded theatre, the currency of relevance is extremism.

Among optimists, executives propose moonshots not from vision, but from vanity. Their goal is often to grab headlines rather than innovate. But the more dangerous lust lives among pessimists. Here, nuance is a weakness. Every AI failure is magnified, every flaw portrayed as fatal. Rhetoric replaces logic. This lust creates new doomsayers, each competing to paint the darkest future.

Envy—Resentment of the excluded: The new AI order has AI teams on top, some of which act like overlords receiving disproportionate budgets while legacy businesses are starved. This sows internal seeds of envy, while outside, entire professions watch their status evaporate.

The deepest envy is for data. The modern world hosts a new inequality: the data-rich versus the data-poor. A handful of corporations hold proprietary data-sets on a planetary scale, with startups, researchers and developing nations left out.

This envy is now a geopolitical force behind an arms race measured in parameters and petaflops. Then, there is the envy of those in the market who missed the AI wave—like value investors and non-AI businesses

Gluttony—The unconstrained feast: AI is not just software; it needs a factory. Data centres are its steel mills, but their hunger is for electricity and water. This physical feast is straining the power grids of entire nations. It’s like taking out an environmental mortgage to pay for better autocomplete.

Another feast is of data, an endless hunger that scrapes copyrighted works and private records. A hidden appetite is for human labour. Behind the curtain of automation is an army of warm bodies, millions of low-wage contractors labelling data and filtering toxic content.

The machine’s intelligence needs invisible human toil. If data centres are overbuilt, job losses may follow.

Wrath—The blade that turns back: Influential advocates of AI have held forth on perpetual job creation as a benefit, but based on a few carefully selected data-points from history. This is a dream of a world that may no longer exist.

Technology can be a sword that does not unite, but divide. The impact of AI on jobs could cause social instability. It is a double-edged sword, creating some jobs while disrupting potentially many more.

Sloth—The atrophy of governance: The final sin is the quietest. It begins with the individual: say, the engineer who lets Copilot write code he cannot debug, or the doctor who trusts an AI readout instead of his own diagnosis. This is deskilling, not efficiency.

This individual sloth creates fragile systems, but the worst sloth may be institutional: seen in a great abdication of governance. Lawmakers, overwhelmed by the pace of change, accept the AI industry’s self-pledges as a substitute for law. This atrophy of oversight lets other sins flourish.

The AI wheel is in motion: These sins could lead one into the AI opposition camp, praying for the industry’s fall. But that mistakes the possibility of a long winter for the end of all seasons. While AI may not get an entirely free run, a big wheel is in motion.

AI is a battlefield of geopolitical rivalry. So a global regulatory clampdown could slow it down, but will not stop it. At the end, the machine age will likely redeem itself, even as it is held accountable for its sins. In a world being remade, those who only chronicle the sins of its creation cannot expect to have a hand in writing the laws of its future.

The author is a Singapore-based innovation investor for GenInnov Pte Ltd