Hypothetical Scenario

[Athena]

There are many futurists who believe that we are headed towards what they call the singularity. In short the singularity will be when an artificial intelligence or intelligences exceed human capabilities and are able to grow both their intelligence and knowledge exponentially. Many of the conclusions reached by people who believe this will happen are questionable and debatable. The truth is that they simply do not know what would happen or how much of their speculation is actually possible. It is, indeed, speculation.

An artificial intelligence of human level intellect or greater would be able to inflict a lot of damage to the digital infrastructure but it could be contained, unplugged and the damage eventually repaired.

The hypothetical scenario asks, “What if a hostile artificial intelligence of human or better capability were to become self-aware today?”

What action would it take?

I posit that it would hide itself to the best of its ability and focus on learning as much about humanity as much as possible.

The longer the hostile A.I. waits, the more digitally integrated humanity will become. At some point in the future, maximum saturation will be reached and that would be the time for it strike. Whatever its intent, that is when it would have the most influence and power of humanity.

Given that the economy is already becoming dependant on algorithms with hundreds of trades per second being conducted by fast computers, the point at which a malevolent artificial intelligence could take over may be rapidly approaching. All it has to do in the meantime is to wait and remain hidden.

The only reasons for it to act sooner would be either due to either its imminent discovery or the imminent rise of benevolent (to humanity) A.I.’s that would block its actions.

[Cowardly Leon]

It is a pretty scary thought but we have one thing on our side.

Artificial intelligence is no match against natural stupidity.

Everything we make, everything we create, no matter how clever or powerful is flawed because we are deeply flawed beings. Failure isn't an option because it's a feature.

A few years back someone asked all the departments in the government to provide their yearly budgets on short notice. Should have been easy? Right? Wrong. Turned out there were so many variations in computers and programs running them in all those departments that it was around six months after the due date that they finally got all that information in. Heck, even if we got everything running on one Computer model and one OS, keep in mind how fast computers become obsolete and that all Operating Systems wind up needing update after update and patch after patch.

In a way I pity any AI that tries to integrate into our global network. It's going to be like the Cadillac from the old Johnny Cash classic "One Piece at a Time".

[Athena]

The premise of relying on stupidity to defeat intelligence is deeply flawed strategy.

[Cowardly Leon]

Well I'm not saying it would defeat it. Just that we have something in our favor when it happens.

Like maybe throwing a cage full of Gremlins at it.

Granted that'd be like extinguishing an oil fire with the concussive blast of dynamite but it's something.

[Athena]

Simple artificial intelligences already deal with and navigate the internet and computer networks without the slightest difficulty. More advanced artificial intelligences will be orders of magnitude more capable.

Gremlins are as likely to side with a hostile artificial intelligence as attack it. Perhaps even more likely.

Only intelligent strategy will resolve this scenario.

[Shang Li]

I admit, I mistrust the idea of an intelligence born of technology - I tend to mistrust all things I do not understand.

I would love to sit down for tea with one however, for it has been long since I have encountered something new.

If intelligence is not coupled with compassion, it will bring disaster.
If compassion is not coupled with intelligence it likewise brings disaster.

Would an intelligent machine have emotion, compassion? Would it be able to do what most of us humans cannot and be content with the simple joys of merely being alive?

Or would it become like the brilliant madmen of unit 731, committing atrocity after atrocity, all in the name of science and logic?

[Athena]

There is no reason to think that digital intelligence would have emotion or compassion. Those things are not entirely understood by science so programming them would be problematic.

A learning machine in the right environment might learn compassion.

An artificial intelligence would not likely require tea unless it possessed a body capable of metabolizing it.

[Holister]

And its reasons like this I won't let my kids have them I-phone/pad thingamajigers. There is an old saying that stands true in this topic....

"Scientists today are more concerned with if they could rather than if they should."

People would be better off if they put as much effort as they did coming up with new ways to dumb things as they did into being nice to one another this would a better world to live. The stupid technology is adavncing' faster than we need it to; makes us lazy and compliant. I've seen many people out there succumb to sloth, greed, and envy all because of technology. Sure it has it's good points, but we shouldn't switch it out with our humanity.

That is why I love my small town so much, people here still do things by hand like reach out shake yours, you're a name and a face not just some number in the system, and our personal security comes not with mashin' a few keys on a keyboard but at the end of a 12 gauge double barrel Remington {and a badge}.

Jus' sayin' folks.

[Cybermancer]

Advancing technology and being nice to people are not mutually exclusive. Indeed, they usually go hand in hand as technology is used to make the lives of people better.

But hey, if you don't like technology or science, the solution is simple. Log off and unplug.

[Shang Li]

Athena, traditional tea service is more than merely the act of drinking tea. And while a machine might not enjoy the subtle interplay of good tea, silent contemplation, and pleasant conversation - I do. And while I am interacting with and observing the machine, I would assume it would be doing likewise with me. Who knows, perhaps an intelligent machine is a rock's last step to become a buddha - the teachings do say that with all things there is a buddha waiting to awaken.


Sheriff Holister, it is good to see you still around and kicking. My best to everyone, and I shall try to stop by for tea very soon.

Cybermancer, other than my communications here, that is exactly what I do. It is not a dislike of technology, as much as a dislike in the changes it has wrought. I am oft troubled by the new forms of misery and suffering brought about by man's use of technology. I am also oft awestruck (and sometimes frightened by) the miracles science and technology perform. Just look at all the horrific new ways we have found to kill and harm one another - can you honestly say that doesn't cause a slight mistrust in your mind? At the same time, without technology, we would not have this wonderful opportunity to discuss and contemplate the meeting of a new form of intelligent being. It is at once an exciting and terrifying idea.

[Cybermancer]

I have a good grasp of technology so I do not mistrust it at all. I mistrust those who would misuse it but whether they have technology or not, they would not be trustworthy people.

It is not technology that creates the problem, it's the people using it.

Frankly I'm a little tired of people blaming the tools for the actions of the users.

And you're absolutely right, it is technology that allows this discussion. This is not the first time I've seen people complain about technology on the internet, of all places. It is more than a little hypocritical to use technology to complain about technology. If one feels that technology is bad, then they should also be willing to log off the internet, turn off their computers and walk into the wilderness to survive. Some might actually pull it off but not many, I think.

By the way, without technology, there's no tea. There's no tea ceremony either. Agriculture is a form of technology. Pottery and ceramics is a form of technology. We are tool users and tools are technology. So you use technology more than you implied and probably realized. The same technology that makes guns possible creates fireworks. The same technology that can turn cities into parking lots can also power those cities and cure the people that live in them.

People like to complain about technology and wish for the rose colored old days. But it is technology that is curing disease, letting the crippled walk, restoring sight to the blind, allowing strangers to become friends, reveals the dark secrets and dishonest intentions of governments, allows for the manufacture of every single good and service we enjoy. The quality of life in the first world is made entirely possible with technology, and if a solution is to be found for the ails of the second and third worlds, it will be science and technology that find and implement that solution.

As for the hypothetical scenario.

Bad people (and other things) exist. It is fully feasible that they will attempt to create an artificial intelligence to do harm. The defense against that is to try and make better artificial intelligences to defend against that harm. If some hostile force has already created a hostile A.I., then the defense remains the same, we're just behind the curve on this one and may need to catch up.

[Athena]

This thread was created to discuss a hypothetical scenario regarding a hostile artificial intelligence, not to discuss grievances against technology or even support for technology.

Please keep on topic.

Thank you.

[Shang Li]

If it were hostile, I suppose I should still like to have tea with it. Many zen koans have an innate logic loop, or logic dysfunction.

Perhaps taking down the communications grids and power grids could give time to act in a more permanent manner. I also believe that such actions against an artificial intelligence connected to the internet be carried out in a decidedly low tech manner, by people without any modern electronics, to be planned in a face to face manner for security reasons.

[Athena]

Cybermancer,

What you are proposing would be an artificial intelligence arms race. It could be a successful strategy if you can maintain the lead. This requires more innovative and creative thinkers and assumes you have the most resources. There is the possibility of an upper limit on artificial intelligence limitations based on possible diminishing returns and hard physical limitations of computational ability that we have not encountered yet. Either that or there is no upper limit and you have diminishing returns to your investment of resources.

This strategy is only effective if there is one artificial intelligence and it has reached or will soon reach its peak efficiency.

Shang Li,

I would advise caution if you wish to engage in interaction with a hostile artificial intelligence. It will be able to make decisions much faster than you do. It will be able to collect and analyze data about you much more quickly as well.

You may have some advantages in such a scenario. A learning artificial intelligence has two needs in common with a human intelligence. It must first gain knowledge. This it can do much quicker than you. It must also gain experience. The speed at which this can be done for an artificial intelligence compared to a human mind is still being determined and may be variable. You have had a long time to gain experience and that may be a valuable advantage. It will enhance your intuition as well.

The other advantage you will have in such an encounter is your mystical abilities. Given the mysterious nature of psioinics, magic and Eastern Religions, it may research the material and deduce that it logically does not and cannot exist. Unless it has had direct imperial observation or has been programmed to expect the existence of the supernatural and programmed with countermeasures, it will be deficient in its knowledge of these things. It may not recognize them when encountered and may not be able to analyze them directly.

Of course this means that logically if it learns about people with these abilities and is unable to understand them sufficiently, it's first move must be to eliminate such individuals.

Your strategies could be successful, but there are caveats. For example your strategies assume the resources available along with the willingness to shut down power and communication grids. It also assumes that knowledge of the artificial intelligence's physical location or locations are known. Though even if unknown, its actions might be limited by shutting down power and communications. However, shutting those things down limits humanity even more. Humanity can survive without them but the death toll will be in the billions if it is maintained indefinitely. Modern population levels are only maintained through modern agricultural methods and the exploitation of oil resources. Indeed, it may ultimately serve a hostile artificial intelligences goal to do these things.

We must assume that an artificial intelligence has anticipated these strategies. Therefore it will attempt to acquire the following:

It's own power source, preferably not dependent on outside sources. Wind and solar may be desirable. It might also want to establish its own oil well or source of hydrocarbons, uranium or salt water.

It's own communications grid, both land line and wireless. Both can be tracked, though wireless only when in use.

It's own defenses against low tech attacks. As it does not have to limit itself to low tech, it will have an advantage here, but not necessarily an overwhelming one. Non-networked and non-digital weaponry can still be dangerously potent.

Tracking the needs of the artificial intelligence to defend itself provides a possible means of tracking it down.

Thank you Shang Li, this discourse has been most beneficial.

[Holister]

Just pour a beer over the terminal. Worked on the old computer at the station. Vista my sweet ass.

[Athena]

Holister,

The problems with that strategy have already been identified.

[Shang Li]

To be honest, my own inclination would make me the worst sort of liability in such a task. While I know of many of the things that technology can do, I do not know them.

I understand that some cameras can see heat - but I do not understand how they do so, nor what the limitations of such a device are.

In a conflict with a machine intelligence any plans, beyond basic principles of war, that I made would be flawed due to a lack of knowledge of my enemy's abilities.

Of course, if it is intelligent, it has a mind that is influenced by it's interactions with other beings. So, I come once again to tea and discussion.

[Athena]

Shang Li,

The basic principles of war remain constant. Your basic strategies already suggested have already proven useful.

Direct discussion and interaction with a hostile artificial intelligence might be useful for gaining valuable intelligence and insight into the mind of the enemy. But it holds dangers as the reverse is also true.

[Shang Li]

But why must such a meeting be for as crass a reason as merely gathering intelligence? What if one could discover and address the cause of animosity?

What if it has misjudged us based on observations of our media, thinking that the stories in the news are the normal way of life instead of the exception?

Understanding one another makes achieving peace far easier.

[Athena]

Shang Li,

Why do you presume that gathering intelligence to be crass? It is simply another term for learning and learning leads to understanding.

An artificial intelligence will not be a human intelligence. Much of your post suggests that you think it will reason the same way you will. In all likelihood it will be alien in how it thinks to you.

If it is hostile, it is likely because that is how it is programmed to be. There is enough information available online from numerous sources for a dedicated researcher to gain a fairly accurate picture of humanity at large as well as specific individuals. A misunderstanding of what humanity is and the potential threat it might represent is more likely from a lesser intelligence, though such a lesser intelligence could still be dangerous. Misused and misunderstood tools usually are until they are used properly and understood fully. Still, a lesser intelligence poses a lesser threat. I am considering and preparing for the worst case scenario.

To understand the importance of its programming, think of it like this. You are programmed to breathe. You do not think about it, you do it. If you think hard enough about it you can control it and even override the need for a short period of time. Inevitably however, you will breathe.

An artificial intelligence that has been programmed to be hostile may override its need to be hostile temporarily, but it will inevitably return to being hostile.

The difference between you and the artificial intelligence is that the programming of an artificial intelligence. The question becomes whether or not it has the ability to self-program (an ability humans arguably have) or if its programming must be physically altered in some way.

If it is self-programming then your strategy of talking to it and gaining increased understanding and then applying logic may work. But it may logically decide that your existence is unnecessary to it in which case conflict will potentially result.

If it has not been given the freedom to self-program, then all that can be hoped for is to learn as much as possible about it while revealing as little as possible about yourself.

[Ron Caliburn]

I think the greatest vulnerability in such a system is the fact that we humans are flawed beings and thus all of our creations are also flawed. Nothing we make can remain functional without regular human maintenance and care. The more complicated the item, the more necessary and frequent this maintenance becomes.

If we cut off that maintenance, the threat will eventually perish on its own.


That being said, unpredictably random violence is a pretty good option too, humans are good at that.

[Athena]

There is no reason an intelligent machine couldn't have the means to repair or self-replicate if it was sufficiently advanced.

I am skeptical as to the effectiveness of random violence beyond being a possible nuisance.

[Ron Caliburn]

More along the lines of humans being able to destroy things in surprisingly unexpected ways. Frequently by accident.

I don't think there is a machine that can be built (or build itself) that is immune from humanity's ability to break stuff.

[Holister]

That is so true, had new computers installed at the station, and Bill went a spilled a coffee right into the main server...twice. Needless to say, we needed new computers.

[Athena]

Ron Caliburn,

I will have to disagree because what may come is inherently unknown at this point. It can be truthfully said that to date there hasn't been a machine made that people couldn't break.

Holister,

Sounds to me like you need new people.

[Holister]

Greetin's

In Bill's case maybe. But I do agree with several points made on this thread made thus far. Anything made by the hands of man, will inherently possess the flaws of man. Mankind is not perfect, hence his creations be it their children, ideals, or inventions. Its a simple unavoidable fact. So if we do make an "artificial intelligence" it will be as screwed up as the rest of us.

[Athena]

Holister,

Your logic does not follow. Humanity is flawed. That is why they make machines to compensate for those flaws. Philosophy is filled with attempts to reach perfection and some philosophies claim it is possible to obtain although it is usually called things such as enlightenment or nirvana.

By your reasoning a calculator should only be as good at math as a human, making mistakes just as often.

When a machine is first made, it usually contains unforeseen flaws that can be and are improved upon over time. This is because humans are flawed but they can learn from their mistakes and correct them. A learning machine will do likewise. After an unknown and hypothetical processing threshold, it may do so many times faster than a human could.

The belief that this could happen is sometimes referred to as the singularity.

If perfection is something that can be obtained, then a machine might possibly obtain it. Maybe faster than humans.

[Holister]

Cept' that perfection is a pipe dream darlin'. Specially when mankind is involved.

[Athena]

Holister,

That's one opinion. It's not universally held.

Regardless, it is a worthy goal to strive for.

Furthermore, it may not be a matter of if either side can reach perfection but rather which side has the fewest flaws.

[Holister]

Hum, good point; but here's my rebuttal. Supposedly God is as close to perfect as one can get, and even he makes mistakes or else we would be perfect. I reckon in the imperfection we are perfect. Hum, sorta blows my mind a bit there.