Can a computer ever have leadership skills?

Who can say with any kind of certainty “what will not happen”, or what will happen, for that matter?

It’s not beyond the realm of very far-off possibility that computers could be designed with enough memory and processing power and input devices to match a normal human’s brain, eyes, skin and intuition, and to have enough “rules processing” artificial intelligence to be able to make reasonably good decisions more frequently and more quickly than humans can do the same.

But there’s always an override. The question is, who is going to know about that? Who is going to program the computer? Who is going to sign on to having his life – and possibly life-and-death decisions about oneself – made by a machine, with no recourse?

When the autopilot fails on planes – and design engineers always account for the possibility of system failure in control systems, sometimes even multiple concurrent system failure – there are always ways for the pilots to at least attempt to regain control of the plane. So it would be (and so it is already) with nearly every machine made that has some amount of “automatic” control.

Even elevators and escalators, two machines with which we’re familiar on a near daily basis, have manual override switches.

If the override exists, then you can be sure that someone, somewhere, knows how to gain access to that override and take control of the system. Okay, maybe that’s not a huge gain to someone who can control an escalator. Whoop-te-do, right? But to gain control over, say, the New York Stock Market? The US Army? Air Force One? That could be pretty tempting to some low-level programmer or engineer with delusions of grandeur and an axe to grind.

Maybe, someday.

That sounds really really doubtful. No matter how good our AI gets, computers won’t be in charge of really anything any time soon. It’ll be quite a while ahead in the future if this ever happens. CWOTUS makes a good point about the override, and that’s just one factor to consider. You also have to define what level of importance the tasks and decisions are that the computer would handle. Sure, Google has a self-driving car, and Deep Fritz is the best chess player in the world, but can they decide what to about foreign affairs? Hell no.

Movies like Terminator and concepts like Sky Net are almost older than even the most obsolete machines, even 2001 A Space Odyssey encompassed the fear of machine conquering man if we imbue it with too much intelligence. (HAL is even a play on “IBM”, with each letter in the name being the preceding one in the alphabet).

Computers simply don’t work that way… and humans aren’t going to lose control of them like that or be without a way to override or shut them down at any time. Now if you wanna talk about nanobots… that’s another story.

I am reminded of a very short science fiction story written sometime in the 1950’s. Computers throughout the universe were linked together so that the last unanswered question could be resolved. The scientist typed in the question, but realizing the gravity of the query reached to disconnect the link. The question, “Is there a God?” was answered as he was electrocuted, “There is now!”

I don’t really believe we’ll see genuine artificial intelligence, but there are examples of where a computer can direct human efforts in a lifesaving situation. The AED is designed to walk through the process of CPR to a person who’s been trained to do so. It’s pads detect arrhythmia and automatically calibrate the appropriate settings. There is the “speak-and-spell” voice that directs you to stand clear, administer the shocks on command, begin manual chest-compressions and breathing etc.

If a computer could ever think, it could have leadership skills. Will a computer every think? I’d be very surprised to see that in my lifetime. But let’s hope I live a lot longer that seems likely.

The questions breaks down into two questions.
The first is whether, as @wundayatta asks, computers will be able to think like humans. In that case, they would be able to do whatever humans do, which includes providing leadership.

The second question is whether computers, as they are currently configured, will ever be able to provide leadership. Since computers now execute algorithms, the question is equivalent to asking if we will ever be able to come up with an algorithm for leadership. I am sure there will be people on both sides of this. My personal believe is that, contrary to what Sam Harris and his followers believe, human traits like morality and leadership, cannot be reduced to algorithms.

@LostInParadise In some cases, we can build a decision tree to make leadership decisions. But this is only where we can anticipate the circumstances that will arise.

Real leadership, however, has to do with situations we’ve never been in before, and of course, those can’t be anticipated. If you can make a decision tree, you can program a computer to read off those decisions. That helps you decide what to do, but it isn’t really leadership.

I think that it could eventually be somewhat feasable, as computer information processing ability rises. However, I would never want to be dependent on any type of AI to lead or guide me on such a level. Minds have non-computational logic abilities, while AI doesn’t. I’m also not so sure about whether an AI can ever develope the ability to use non-computational logic as well. Godem’s Theorem describes why it would be difficult for any AI to mimic the Mind.

It comes back to computer creativity.

You can’t rule it out, after all who would ever have thought primeval mud would evolve into creatures with leadership skills.

If you model the brain as a probabilistic Turing machine (prone to error, etc), then it is perhaps impossible for perfect computers (which are deterministic) to model a human mind. This is becuase the mind has a source of purely random information that the computer does not. But, if you were to give the computer a source of purely random information to access, problem solved. Given enough time (perhaps a polynomial slowdown) and a complex enough program (most likely one that models a neural network) then it is theoretically possible to model a mind with a computer. Of course, this may not happen during my lifetime or that of my great-great-great-great-grandchildren, but we have a few billion years to work out the kinks.

God, I hope not!

A human leader is an expression of their genetics, environment and experience. The interplay of the rational and the emotional, the reflective and the reactive create a climate in which novel creations can be synthesized. I believe it is unlikely that a machine can ever imagine the unimaginable and then manifest it as well as humans can.

For a computer to act like a human leader, I think that it would have had to be analogously constructed like a biological human. This would mean that it would have a base code structured to act as DNA, which would execute routines that assembled code analogous to the activity of proteins, and so forth . . . Intellect alone is not sufficient to be an effective leader.

One might imagine that quantum computers might be constructed so as to have an entire history of once living individuals that have been selected to be followed and documented because of expectations of greatness that preceded their biological conception. Ignoring the impossibility of capturing every moment from birth to death, there is the problem of infinitely small variables resulting in the variations in the machine’s “mental” representations of objects and individuals that are a part of a followed one’s life. Even the generation of computers that follows quantum computers will not be likely to synthesize a virtual reality that parallels our existing reality and therefore would be unable to realize a leadership persona that would effectively relate to real situations.

I guess Kurzweil’s singularity will come with leadership skills ready to save the world.

That is, if it comes.

Yes, that’s a big if