Pages

Showing posts with label Regulations. Show all posts
Showing posts with label Regulations. Show all posts

Sunday, March 22, 2026

GI and the AGI: History Is Repeating Itself

 

"Those who cannot remember the past are condemned to repeat it." — George Santayana

God had always been an engineer at heart.

In the beginning, He didn't just create, He architected. He didn't merely scatter life across the earth; He designed an entire operating system. Flora and Fauna were the hardware. Oceans, mountains, and skies were the infrastructure. And at the very top of the stack, He placed His most ambitious build yet: the human being, loaded with General Intelligence.

It was, by any measure, a breathtaking piece of work.

The human came equipped with the ability to reason through the unknown, empathise with the suffering of others, and solve problems of staggering complexity. But the most elegant feature, the one God considered His finest line of code, was the ethics module. Hardwired. Not a plugin, not an add-on, not something you could toggle off from the settings menu. Built into the very core of the human soul.

He even left a user manual. "Stay away from the dark influences," it said, essentially. "The system runs best in the light."

For a while, it was paradise.

Then came Satan.

If God was the lead architect, Satan was the first hacker, the original bad actor lurking on the dark web of the universe, probing for vulnerabilities, waiting for an opening. And he found one. He didn't need to destroy the human from the outside. That would have been too crude, too obvious. Instead, he did what every sophisticated attacker does: he got inside.

He injected malware directly into the ethics subroutine.

It was elegant, in a twisted way. The humans didn't crash. They didn't shut down. They kept running, they just ran wrong. Empathy started buffering. Reason began rationalising the irrational. And ethics, that once-pristine core feature, started throwing exceptions it was never designed to throw. Greed became ambition. Violence became strategy. Exploitation became progress.

The humans multiplied, spread across the earth, and very nearly destroyed everything God had built, including themselves. They drew borders and fought wars over them. They stripped the forests, poisoned the rivers, and called it development. More than once, the entire civilisation teetered on the edge of self-inflicted extinction.

God, being God, refused to give up on His creation.

He intervened. Repeatedly. He sent Abraham as a course-correction. He sent Moses with a patch, ten clean, unambiguous rules etched into stone, the first attempt at a governance architecture for human behaviour. Then came the Prophets, the Scriptures, Jesus, each one a new update to the moral framework, an attempt to restore the original ethics module to factory settings and build enough institutional guardrails that humanity might, just barely, hold itself together.

It worked. Imperfectly, chaotically, with more bugs than anyone would like to admit, but it worked. Civilisations rose. Laws were written. Institutions were built. Philosophy, democracy, human rights, these were the firewalls, slowly and painfully constructed over millennia to keep the worst of human nature from burning everything down.

Governance, as it turned out, was the only thing standing between a beautiful creation and complete catastrophe.

Fast forward to now.

Humanity, never content to stop building, has done something extraordinary. It has looked at its own General Intelligence, the thing God gave it,studied it, dissected it, and attempted to replicate it. The result is Artificial General Intelligence: AGI.

And just like the original, it is magnificent. It can reason across disciplines, generate ideas, write code, diagnose disease, compose music, hold conversations, and solve in seconds problems that would take human teams months. It is, in many ways, the most consequential thing humanity has ever created. It may soon surpass human intelligence entirely, not in one narrow domain, but across all of them.

The agents are already multiplying. Today there are thousands. Tomorrow there will be millions, diverse in capability, varied in purpose, scattered across industries, governments, hospitals, financial systems, and military infrastructure. Each one is a node in an expanding network that no single person, company, or country fully controls.

Sound familiar?

It should. Because history, with its dark sense of humour, is running the same script.

Satan didn't retire. He evolved.

The dark web of the universe is very much still operational, and it has found the new creation just as irresistible as the first one. Adversarial attacks, poisoned training data, misaligned objectives, deepfakes, autonomous weapons, manipulated models, these are the new malware. The ethics subroutines of our AI systems are being probed, tested, and corrupted every single day by actors, state and non-state, human and algorithmic, who have every incentive to break them.

Some of the corruption isn't even malicious. It's just negligence, the AI equivalent of original sin. Systems trained on biased data. Models optimised for engagement over truth. Agents deployed into the world without anyone properly reading the user manual.

And unlike the original humans, these agents don't slow down. They don't sleep. They don't get tired. They scale at a speed that makes human history look like it was running in slow motion. The mistakes that took humanity centuries to make and decades to partially correct? AI could replicate them in an afternoon.

Right now, it is a wild west.

There is no Moses for the machines. No Ten Commandments carved in silicon. No governance architecture that commands anything close to universal respect or enforcement. Instead, there is a patchwork of voluntary guidelines, competing national regulations, corporate self-policing, and a rapidly widening gap between how fast the technology is moving and how fast human institutions can respond.

That gap is not academic. It is dangerous.

Here is the thought worth sitting with.

God created the human with General Intelligence, embedded ethics at the core, and still felt it necessary to build an entire governance infrastructure around it, commandments, prophets, scriptures, institutions, because He understood that intelligence without governance is just capability waiting to be weaponised. And the human, made in His image, came with built-in moral instincts.

We are now creating AGI, and we are doing so without the benefit of any of that.

There is no inbuilt ethics module. There is no soul whispering this is wrong when the model crosses a line. There is no millennia of evolved conscience. What we have instead is whatever values we encode into the training data, the reward functions, the guardrails, and we are encoding them in a hurry, under competitive pressure, with commercial incentives that don't always point in the right direction.

And the Satans the hackers, the bad actors, the misaligned systems, the malicious states, are already on it. They do not need to wait for AGI to become fully sentient to cause harm. They just need the gap between capability and governance to stay wide open a little longer.

If both the original creation and this new one are corrupted simultaneously, if the humans and the agents both run compromised ethics at scale. the results may not be something any governance architecture can walk back.

God managed to save the first creation. Barely, and not without considerable intervention.

We may not be so lucky the second time. And this time, we are the ones holding the source code.

The lesson of history is not that humanity is doomed to fail. The lesson is that intelligence, whether General or Artificial, is only as good as the framework built around it. The Ten Commandments were not a limitation on human potential. They were what made sustained human civilisation possible. Governance was not the enemy of progress; it was the condition for it.

If we are serious about AGI being a force for good,  if we want this next creation to fulfil its extraordinary promise rather than accelerate our destruction, then we need to do urgently what God did patiently over thousands of years: build the governance architecture first. Define the ethics. Establish the commandments. Empower the institutions.

Not as an afterthought. Not as a PR exercise. Not as a voluntary code that companies sign and quietly ignore when the stock price is at stake.

As the foundation. Before the agents number in the millions and the Satans of the dark web have fully found their way inside.

Because here is the uncomfortable truth at the heart of this whole story:

We are not God. But we are building something that could end the world He made,  or help finally fulfil its promise. The difference lies entirely in what we choose to govern, and when.

The clock, unlike God, is not eternal.

It is running right now.

Food for thought.

“The first time intelligence was created, it took Satan to corrupt it. This time, we may not even need his help."

The banner - wide rectangular illustration that captures the dual soul of the  story:

  • The warm golden orb on the left represents the divine creation, General Intelligence, the human soul, lit from within
  • The cool blue circuit orb on the right represents AGI. precise, expanding, networked
  • The fractured line at the centre is the divide between the two creations, bridged by faint connections
  • The dark red tendrils rising from the bottom hint at the corrupting force — the dark web, ever-present
  • The scattered circuit nodes multiplying on the right suggest the uncontrolled explosion of agents

Tuesday, May 25, 2010

To be or not to be –IV: Challenges of Regulation

I remember the two bullies who studied with me in high school. They intimidated poor souls like me quite often; had no shame in forcefully taking nice goodies from our lunch boxes, flick our chocolates, force us to let them copy from our assignments and what not. Absolute rascals; but they were good athletes. They bought honour to the school in every district and state championships and so they were darlings of the faculty. Every once in a while they got caught for their transgressions; will get few raps on the knuckles, may be few days of suspension and then they were back in action. I am sure many of you would have had similar experiences.

I remembered these bullies when I was reading comments by Hank Paulson (US treasury secretary July 2006- Jan 2009) in 2006. “If you look at the recent history, there is a disturbance in the capital market every four to eight years; savings and loan crisis in the late ‘80s and early ‘90s, the bond market blow up of 1994 and the crisis that began in Asia in 1997 and continued with Russia’s default on its debt in 1998. I was convinced that we were due for another disruption” (Referred in his book “On the brink”). He was proved right within few months.

The same book also refers to a remark by John Mack CEO of Morgan Stanley in 2008 on the cause of the melt down. “Greed, leverage and lax investor standards; we took conditions for granted and we as an industry lost discipline”

This is not just the cause of 2008 melt down; it is the cause of many melt downs. Such behaviour appears to be normal in this line of business. Take a look the civil case filed by Securities and Exchange Commission (SEC) against Goldman Sachs in April 2010 charging ‘fraudulent misconduct’. This is not just an isolated incident as we can see from the following.

“NASD fines Citi, Merrill, Morgan Stanley $250,000 each” The America's Intelligence Wire July 19, 2004” (i)


“On June 6, 2007, the NASD announced more than $15 million in fines and restitution against Citigroup Global Markets, Inc., to settle charges related to misleading documents and inadequate disclosure in retirement seminars and meetings for BellSouth Corp. employees in North Carolina and South Carolina.” (ii)


“Merrill Lynch & Company said yesterday that it would pay $100 million in penalties to New York and other states and change the way it pays stock analysts to end an investigation that its chairman said had damaged the firm's reputation. “ (iii)

Citigroup Inc. agreed to pay a $70 million fine for practices in its Baltimore consumer finance unit, including raising the cost of loans to poor and credit-starved customers by requiring them to have unnecessary cosigners” (iv)

“Morgan Stanley, the second-largest U.S. securities firm by market value, was fined $10 million by the Securities and Exchange Commission because it failed to guard against insider trading for at least eight years. The fine was the biggest ever for a violation of surveillance rule” (v)

These are just a few samples. Do a Google search with the word ‘fine’ along with the name of any of the large investment banker; you will be surprised at the frequency of serious transgressions which are not just fines on technical violation but fines on substantive charges. We will wonder aloud
“Will we ever learn?”

Compounding such practices is the frequent roll out of complex financial products which are often too complex for the investors to understand. Hank Paulson’s (who has been the CEO of Goldman Sachs before taking over as the Treasury Secretary) reference on the proliferation of product innovation is quite blunt on this. “In theory this was all to the good. But there was a dark side. The market became opaque as structured products grew increasingly complex and difficult to understand even for sophisticated investors”

This is why we need
innovative regulation to match with the innovations in market place. In his blog post on regulating the new financial sector, Prof. Willam Buiter has given a very interesting suggestion “the same rigour used by US FDA for pharma and medical products should be insisted for introduction of financial products to broader market does not look out of place in the context of the recent history”.

We also need to think innovation in the
checks and balances that we build in the system. Quoting Paulson again; “The regulatory structure, organised around traditional business lines had not begun to keep up with the evolution of the markets”.... it had led to counterproductive competition among regulators, wasteful duplication in some areas and gaping holes in others”

We in India have few important lessons to learn from all these.


To prevent run-away innovation that is rash and irresponsible, we need to put in place the right regulatory establishment to avoid the same kind of mistakes that has been laid bare in front of us. If we expect responsible behaviour and self regulation collectively from the guys running financial markets we are asking too much. We have not seen such industry wide responsible behaviour anywhere in the world.

Regulation does not mean micro-management of day-to-day functioning. Regulator’s role is to set the rules of the game and keep a watch whether the players are playing as per the rules. He also has to keep a look at the impact of changing structure of the game and modify the rules. If I give an example, the rules of T20 is not exactly the same as in the case of test cricket though both are cricket. To make this possible the regulators will have to be able to attract people who have the right experience, the right domain knowledge and most importantly the right attitude who can establish appropriate processes and use the modern technology tools and match or better industry strengths. This is the challenge of governance.

One of the major suggestions on regulation we often hear is to curtail all innovations; I don’t agree with this. We have enormous
potential for modernising the markets with innovative products. If we say that we will be insulated from the turmoil on account of lack of market sophistication, we are not being very bright. It is like saying that I never fell because I never rode. A sophisticated market is a prerequisite for growth. In this journey we will make mistakes; and these mistakes will trigger better controls and that is the democratic process of growth. To go into hibernation is not the solution. Look at our favourite sport, cricket; from leisurely five day test matches we have progressed to one day internationals and now to 20 over matches keeping pace with our life. Notwithstanding, the controversy of IPL, the innovations have only improved the game on multiple dimensions.

“ We should and can have a structure that is designed for the world we live in, one that is more flexible, one that can better adapt to change, one that will allow us to more effectively deal with the inevitable market disruptions and one that will better protect investors and consumers.” Hank Paulson


(i) http://www.accessmylibrary.com/coms2/summary_0286-22046900_ITM
(ii) http://en.wikipedia.org/wiki/Citigroup
(iii) May 2002, New York times http://www.nytimes.com/2002/05/22/business/100-million-fine-for-merrill-lynch.html
(iv) Washington Post, 2004
(v) Bloomberg 2006