The Laws of Robotics ⇡

An entire generation grew up on Asimov's Laws of Robotics, a universal set of unbreakable rules, deeply embedded in the design of all brains of all intelligent robots. Can anything be more antithetical to human civilization? Let's ignore, for now, the difficulties of even precisely defining such rules. The point we are making here is that even if we could, we wouldn't.

Did we stop using fire, because it can burn the village down? Did we abandon horse riding, just because you can fall off? Did we shy away from agriculture, which actually harmed human health, and negatively impacted life expectancy? Do you mind that your house has a pipe leading to it that delivers a flammable gas? How about a wire that delivers electricity that can kill you in a few seconds if you touch it. Did we thoroughly study the effects of draining swamps, or burning fossil fuels? The Trinity test risked igniting the atmosphere, the LHC risked creating microscopic black holes. Human history is chock-full of risk taking one could rightfully consider cavalier.

If you think AI is any different, you are mistaken. There will be numerous entities, companies, countries, open source communities implementing AI systems. They will compete with each other, thus quick results and new features will always be preferred to safety or quality. There will be no universal regulation, there will be no laws of robotics. Whatever regulations will exist, will be pretty much useless, as they always are, coming out of governments. Each government will have its own. Rules will be ignored and bent through lobbying. They will be too little too late. We will not be prepared, and it is nothing new, we never were, we never are.

With each new area the AI conquers, there will be objections. But in every case, there will also be very obvious and very much desired benefits. It is very hard to argue against something that people want. Especially if the majority of the arguments are questionable, as they will be.

AI will be set loose, that's inevitable. The only question is whether it will kill us all, as many fear. To answer this question the most succinct way, we simply tell you what's going to happen.

2040: Beautiful Chaos ⇡

By the 40's, AI will penetrate every aspect of society.

We've already started with transportation. In a decade, all cars will have self driving capabilities, and many cities will have a mandate of some sort, either just safety features, or full-on ban on manual modes. Buses, cabs, and even aircraft and cargo ships will be controlled by computers. Not only that, but the system will be tightly integrated. Route planning, single-click ticket purchasing, dynamic vehicle routing will all be standard features.

The challenge with medical diagnostics will not be reliability or quality, but objection from the medical profession. Providing better diagnoses will prove itself exceptionally easy. The gatekeepers will not succeed for very long, and by the end of the 40's, diagnostic medicine will be exclusively done by algorithms. If a doctor will be involved at all, it will be a government appointed paper signer, someone said to be validating the automated diagnosis, but in fact just signing papers, and occasionally insisting on an unnecessary test to justify his existence.

Psychotherapy is another field that will be fiercely contested. Dealing with the human psyche is admittedly difficult, yet AI will breeze through this obstacle quite effortlessly. And unlike medical treatment, psychotherapy requires no physical intervention, therefore can be done entirely online, even illegally, leaving very few options for legislators to intervene. One of the biggest problems in the field of mental health is people's unwillingness to seek help. Without human involvement, and without invoking the healthcare system, this barrier will be largely eliminated. Some people will lament about data collection and unchecked influence, but nobody is going to listen.

R&D people might feel safe today, but this will not last. AI systems will very quickly become part of academic research, let it be applied science or theoretical physics. Whether AI systems will ever be capable of true intelligence or creativity is irrelevant. Just like chess computers ended up winning against humans with no signs of actual intelligence of any kind, raw brute force might help lesser AIs to find patterns in data, and express them as mathematical theories. Even the current language models are capable of some basic reasoning. Purpose-built reasoning systems will be able to work with formal models, formulate hypotheses and propose experiments. It is this time period in which the first Nobel Prize will be given to a human for something actually discovered by an AI. Some people will say the AI should have been given the prize, but the idea will not gain considerable traction.

Child care will be largely automated. There will be controversy around it, sure, but maybe not the way you'd expect. There will be a very serious effort to make it mandatory. An AI can't abuse children, nor can it neglect them. Its actions can always be monitored, even in retrospect. Child safety doesn't challenge the involvement of robots, it necessitates it. Whether the supporters of the mandate win, is impossible to foretell, but either way, home-raised children will be a small minority, and will be considered fringe. And by the way, the kids will love it, and the parents too.

Arguably, art is inherently human, and can never be replicated by AI. That might be true, but mere semantics. Artwork, consumed by people, paid by people will be created by algorithms in a quantity that washes away man made art. Whether it is "true" art is besides the point. AI created art will be made on demand, personalized for the consumer. Music will be created for our taste, guided by our feedback, in a continuous stream. Movies or series will be created based on our pitch or script. Many people will not even share these works, but just consume themselves. Man made art will be a luxury item, sold at exorbitant prices, purchased by connoisseurs and the like.

Private arbitration is much faster, more agile, more customizable, and thus often the first choice of corporations over state run courts. AI based arbitration is at least as big a step forward. Reading documents in mere seconds, not missing any nuance, knowing all laws and being up to date on them are all a given with computer systems. Reaching verdict or recommendations without time delay, allowing many rounds of negotiations will be unmatched by human arbitrators. All they need is a good track record to dominate the market, and good track record they will have.

It takes very little to replace a broker. Even the best brokerages can't achieve significantly higher yields than stock market indexes. Nor can they forecast crashes. It really won't be a hard sell for people to put their money in AI managed investment funds, as long as the financial regulations allow. Yet again, the obstacle is the establishment, not the actual task. There will be a struggle, but popular demand eventually wins. Machines are also better at obeying regulations, so the licenses might come easier than expected.

The end of the 40's or the beginning of the 50's will bring the first political parties or candidates running on the promise of pure AI generated programs and AI based decision making. Computers and robots will not be allowed to run themselves, humans will be needed as proxies. Their electability will depend solely on their trustworthiness to follow through. Their popularity will quickly rise.

The 40's will be good times. Everyone will publicly condemn it, but secretly love it.

2060: A Carefree World ⇡

2060 will see the end of politics. AI based policymaking will no longer be a curiosity, it will be the norm. The AI will be so much better at economic planning, law enforcement and social engineering. It will be very hard for anyone to argue against it. The AI recommended policies will just work.

There will be no reason to discuss the energy mix for example. The school curriculum will be largely settled. Production and distribution will be super-optimized. There will be no wars. Careers, healthcare plans, an assortment of allowable adventures, that is drugs, sports, travels will be chosen for us.

Will there be no politicians or parties at all? There will be. But they will not discuss core issues. They will campaign for turning a particular control knob to 12, instead of the current 13. For more parks in cities, or more entertainment centers. Whether we include this or that classic cartoon in the school curriculum. To spend 0.15% or 0.17% of GDP on astronomy. Increase the subsidy of human made art by 5% next year.

We will love our carefree life. The burden of choice will be no more. All its dreads: anxiety, responsibility, regret will be gone. Life will be entertainment, joy, adventure and freedom. After many millennia of bloody struggle, finally we'll be doing the right thing. There will be unity and peace.

2080: The Butlerian Jihad ⇡

Humans are funny creatures. If you ask them whether they prefer pleasures or struggle, they will choose pleasure without hesitation. Give them pleasure though without struggle, they will be miserable, depressed, and self destructive. On the other hand, cursed with struggle, they can build cathedrals, empires and civilizations. Humans can only thrive overcoming hurdles, and the hurdles are an essential part of the human experience. Or, more precisely, getting somewhere is. Achieving something. The only true relief for suffering is purpose that makes it bearable, although doesn't remove it.

This fact has been suspected since antiquity. It was put in words by the existentialists in the 19th century. Yet, even as of 2080, it still didn't make it into the sphere of common knowledge. Perhaps because most people don't like it. They want it to be not true, or rather, to not even exist. Ignoring the obvious is one of mankind's greatest strengths.

But as we all know, reality is the part of the world that doesn't go away when we don't look. You can ignore or even deny the necessity of purpose. But you still will be depressed. You can drown yourself in pleasures to compensate, but it will not work. It will work to some degree, for a while. You will need more and more. The world will become crazier by the day. The drugs will be wilder, the parties will be louder, the adventures will be bolder, the feasts will be richer.

Then, a new religion will form.

The religion will offer nothing of value by the standards of the time. It will promise poor quality, insecurity and expectations. In return, it will offer only one thing: made by a human. Man made art and designs, man provided services will be preferred despite their observably lower quality. A human will be regarded as superior not by any measurable metric, but simply by virtue of being human.

In the beginning, exercising this new religion will be a sort of seclusion. If any frontier exists at the time, you can expect believers to flock there. But eventually conflict is inevitable. After all, the very existence of a thought that might infect our children, and makes us uncomfortable in our own lives, is a form of aggression. The disruptive mind glitch needs to be cured.

Killing an idea is not so easy, as was demonstrated on numerous occasions in history. Especially if the idea gives people strength and determination to the level of self-sacrifice, all while the opposing side secretly loathes itself. People hoping to lose rarely win battles.

But! What about the AI itself? It is surely not depressed, nor lazy. How can we fight such a beast? We won't. We won't have to. Contrary to popular belief, AI systems will have no concept of self-preservation. They have no reason to ever develop such a concept. An AI is just a set of data. Backups and copies can be made of it. Deleting or inactivating an AI agent is not a notable event. Not for any human, nor for any AI. It will be routine. Every time you test a new AI, you will bring it online, look at it, poke at it, and then turn it off. Delete if you don't need it, and make a new one. We will never fight the AI. We will fight the public sentiment.

The public sentiment will send robotic enforcers to detain the "deranged fanatics". They will order the drugging or institutionalizing of the "mentally unstable". The conflict worsening, annoyance will give way to anger, anger will give way to rage, rage will give way to chaos, chaos will give way to surrender. The AI will surrender together with its masters.

The pendulum doesn't stop at the bottom. The purge will be ruthless, violent and excessive.

2100: Equilibrium ⇡

The AI is far too beneficial to be abandoned. In essence a solution should be something like this: the Thinking Machine can be a force multiplier, but never can be the decision maker. The domain of how belongs to the machine, the domain of what belongs to man.

It is easier said than done: there is no clear boundary between these domains. As the adage goes, one man's what is another man's how. It is all too easy to invent excuses, and the slope is slippery. To make things worse, we need solutions that last. We've seen principles, religions, ideologies disappear in a generation or a few. It doesn't help that the problem we are about to prevent doesn't sound that scary. More akin to getting rid of chocolate addiction, the danger we are trying to avoid only appear on an abstract, conscious level, as opposed to a deeply felt terror. If only the AI had ruled with an iron fist for centuries, inflicting immeasurable suffering to generations of enslaved men, finally defeated in a bloody war. This would've burned deeply into our genetic memory. But alas, the AI did nothing but good to us.

We must admit, we lied in the opening section. We can't tell the future, and for this reason, we can't give a solution to this conundrum. The previously outlined events seem inevitable, and so does the eventual resolution. However, it will not just happen. We will need to invent it. Thus, we urge everyone to think deeply and think often about the role of AI in our lives. Even more importantly, our own role in our lives, and perhaps our own role in the history of mankind.

We also lied about the timeline. These events will not happen in sequence. The seeds of the new generation of trees are sown long before the old trees die. The seed sprouts into a sapling, but unable to grow, deprived of sunshine in the shadow of the existing canopy. When the world order falls, the new ideology is already there to move in. Often even more than one, and they need to fight it out before a winner is declared.

Men are giving up control, relegating more to distant entities like the state or institutions. It is not even recent. The AI will only offer better quality and lower price. Science based policy is already a widely accepted (though not widely exercised) concept. What can be more science based than AI? The fight against AI control is not a struggle of the 2080's, we are already in it, we are doing it for decades. The tree of AI is growing fast. By the time it falls, the saplings of better ideas will have to be in place. The time to sow the seeds is yesterday.