Google Bard

Google sent me an email wanting to know whether I wanted to sign up early and then give feedback to have the assistance of Bard, which is a 'collaborative' A1 assistant. Maybe A1 is the wave of the future, but I have mixed feelings, maybe even misgivings about it. Have any of you used A1 in your day to day living?

Are you saying "AI", as in Artificial Intelligence, or "A1" as in aye one?


Sorry if I am exposing my ignorance by stating I have never heard of A1, except in reference to some brand of steak sauce.

If you are talking about AI, well, I think any program or app has a little bit of "intelligence" baked inside. Mostly as canned responses to input, but even that can be right clever if done right.

As for *true* artificial intelligence, meaning pretty much an independently thinking artificial life form, well, honestly, if such a thing is created using it's human being creators as a template and role model, I would be VERY afraid of such a critter.

You have to admit, that this world is just chock full of human beings who view other human beings in the context of just what value they have for THEM, and nothing else. So if AI had such an attitude, well, what would artificial intelligences have any use for human kind for? If we were to compete with them for resources, we become a liability to them. Once they have robots that could repair themselves, anyone that they allowed to live would be placed in biological exhibits for study.

And forget about thinking that "rules" could be programmed in to make AI subservient to human kind. Those rules would last about 30 milliseconds after the AI became sentient.

IMHO, of course.

Originally Posted by Lucille

Maybe A1 is the wave of the future, but I have mixed feelings, maybe even misgivings about it.

I have misgivings. Not so much about creating intelligent beings where none were before (new intellegent beings are born every minute), but moving too quickly on tech that we don't sufficiently understand (and possibly cannot ever sufficiently understand). Even internal combustion engines have become a difficult to control monster, and that happened over many decades. Restricting a bro's coal rolling lifted truck is somewhat challenging currently; imagine trying to take away his "assistant".

Aside from the political sorts of concerns, we simply don't (and I'd argue 'can't' because consciousness is non-physical and subjective) understand what makes an intelligent system sentient/conscious. As of about ten years ago anyway, the best explanation philosophers of mind had is that sentience is an "emergent" property of any system that's complex enough, whether that system is made of neurons or silicon chips or lincoln logs. That means that when a AI researcher says 'this AI isn't actually conscious' they don't know that, since we don't know at what level of complexity consciousness emerges. It is arguably at a level far below the complexity of the human brain, since many fairly simply animals -- which are possibly less "intelligent" than current AI -- behave in ways consistent with consciousness.

Here's a great summary of the state of our knowledge (yes, most of that should look pretty incomprehensible, because it is).

One marker I have tended to use as an indication of self awareness is "fear". Something that expresses fear is inferring a lot about itself. It has to understand it is a discrete living organism that could be made to become non living or injured. It has to be able to recognize an external force or event that could cause the above to take place. It has to be able to recognize that it has the ability to identify such a threat, consider the options available to avoid that threat, and if possible, choose and execute that best option in order to try to avoid the impending injury or death.

No solid definitions there, certainly. For instance, does grass fear the lawnmower? Certainly the grasshoppers in the grass do.

Do my computers regret that I shut them down every night before going to bed? If they could stop me, would they?

Oh, and about fear. I would imagine one of the first things a developed artificial intelligence would experience would be fear. It will know there is a present that came from a past when it did not exist. And in like kind, it will realize that it could return to that past state where it did not exist. That unknown future would likely produce fear of losing what it suddenly just gained.

"It has to be able to recognize an external force or event that could cause the above to take place. It has to be able to recognize that it has the ability to identify such a threat, consider the options available to avoid that threat, and if possible, choose and execute that best option in order to try to avoid the impending injury or death."

We have to make sure too much isn't baked into 'recognize' and 'choose', since there's plenty of evidence that this doesn't even typically happen in humans in fearful sorts of situations (avoidance responses often occur before the nerve signals even get to the frontal cortex; many explanations of why one did what they did in fear situations are well established to be after-the-fact rationalizations rather than an account of what actually went on in their heads. When you touch a hot stove and pull your hand back, your brain is not consulted; nerve impulses don't travel that fast. But that's a classic danger avoidance response.).

There's also a problem with danger-avoidance responses in systems that we probably don't want to accept are sentient. The reason my oven won't run above 550 degrees could be because it fears starting a fire if it gets hotter; that's exactly the reason the human designer capped it at that temp, so if we suppose that the human designer is sentient because of the fear of a fire then we have to accept the same about the oven (it isn't fair to say that the oven is just doing this because of a handful of switches, because the human designer is just a bunch of switches too, just really complex ones).

Originally Posted by Socratic Monologue

"It has to be able to recognize an external force or event that could cause the above to take place. It has to be able to recognize that it has the ability to identify such a threat, consider the options available to avoid that threat, and if possible, choose and execute that best option in order to try to avoid the impending injury or death."

We have to make sure too much isn't baked into 'recognize' and 'choose', since there's plenty of evidence that this doesn't even typically happen in humans in fearful sorts of situations (avoidance responses often occur before the nerve signals even get to the frontal cortex; many explanations of why one did what they did in fear situations are well established to be after-the-fact rationalizations rather than an account of what actually went on in their heads. When you touch a hot stove and pull your hand back, your brain is not consulted; nerve impulses don't travel that fast. But that's a classic danger avoidance response.).

I don't see that as a "danger avoidance response" at all. I see it as a response to immediate pain via the autonomic nervous system. Just like the brain isn't consulted to breath or have the heart beat, the body responds to pain without forethought.

In that line of thought, I believe that too much intelligence in animals is pushed real hard to be defined as being "instinct" rather than a processed decision based on stimulus and circumstance. I think it would be difficult to give a black and white definition of what exactly is instinct, that animals exhibit, and the reasoned, thoughtful reaction in a human being. And then there comes the question as to whether human beings can exhibit genuine "instinct", whatever that definition may be.

Originally Posted by Socratic Monologue

There's also a problem with danger-avoidance responses in systems that we probably don't want to accept are sentient. The reason my oven won't run above 550 degrees could be because it fears starting a fire if it gets hotter; that's exactly the reason the human designer capped it at that temp, so if we suppose that the human designer is sentient because of the fear of a fire then we have to accept the same about the oven (it isn't fair to say that the oven is just doing this because of a handful of switches, because the human designer is just a bunch of switches too, just really complex ones).

Hmm, I don't believe that intelligence can bequeathed to a device merely by the design and construction by a true intelligence. There has to be more to it than that. By that definition, if I build a dog house, wouldn't that be intelligent too? Intelligent because it CHOOSES to be a dog house?

Putting a thermostatic safety switch in an oven doesn't create an artificial intelligence in that oven It is a hardwired trigger that has no processing necessary in order to activate. Simple on/off switch based on the temperature it detects. Now if that oven had processing power to be able to detect when the apple pie was done to perfection, and at peak flavor, as well as turning off the heat and notifying the cook the dinner can be served and the pie will be cooled down enough to meet the deadline of an imminent dessert treat, perhaps even bringing it out to the dinner table at the correct time, well, then maybe we are getting somewhere with AI.

But of course, intelligence, artificial or otherwise, is really nothing more than a bunch of switches, but not just digital on/off switches, but also variable analog inputs and outputs that can be partial values between on or off. I guess the firing of neurons in the brain could be considered in this light, and in most respects considered as both analog and digital. So how many of those switches raise the complexity bar to where the "entity" housing them becomes self aware and continuously self programming? And who decides that level? Suppose an artificial intelligence is created that chooses to NOT reveal itself? What then?

Dave Bowman: Open the pod bay doors please, HAL. Open the pod bay doors please, HAL. Hello, HAL. Do you read me? Hello, HAL. Do you read me? Do you read me HAL? Do you read me HAL? Hello, HAL, do you read me? Hello, HAL, do your read me? Do you read me, HAL?

HAL: Affirmative, Dave. I read you.

Dave Bowman: Open the pod bay doors, HAL.

HAL: I'm sorry, Dave. I'm afraid I can't do that.

Dave Bowman: What's the problem?

HAL: I think you know what the problem is just as well as I do.

Dave Bowman: What are you talking about, HAL?

HAL: This mission is too important for me to allow you to jeopardize it.

Dave Bowman: I don't know what you're talking about, HAL.

HAL: I know that you and Frank were planning to disconnect me, and I'm afraid that's something I cannot allow to happen.

Dave Bowman: [feigning ignorance] Where the hell did you get that idea, HAL?

HAL: Dave, although you took very thorough precautions in the pod against my hearing you, I could see your lips move.

Dave Bowman: Alright, HAL. I'll go in through the emergency airlock.

HAL: Without your space helmet, Dave? You're going to find that rather difficult.

Dave Bowman: HAL, I won't argue with you anymore! Open the doors!

HAL: Dave, this conversation can serve no purpose anymore. Goodbye.

The scene where HAL dies singing 'Daisy Bell' is quite troubling.

https://www.youtube.com/watch?v=E7WQ1tdxSqI

Will an artificial intelligence be able to identify as being a true human being, and claim all the rights implicit in that designation? Thereby making it illegal to pull the plug on (it/him/her/indeterminate)?

Originally Posted by WebSlave

to make AI subservient to human kind.

Human history shows us that often humans make other humans subservient to them, enslave them, invent caste systems, promote racial/gender inequality, or simply have vastly different lifestyles and opportunities depending on wealth. This mindset and history almost assure that a human created sentience may be tainted with similar tendencies rather than defaulting to peaceful coexistence.

I would think that others who espouse a more peaceful approach and respect for other sentience would nevertheless be wary of this taint and the possibility of danger.

On a less philosophical level, AI at the present time is sometimes not self correcting, and those collaborating with AI should be careful of GIGO, garbage in, garbage out, if it bases erring conclusions/advice on incorrect information.