[frankinstien]

I don't know how often I have had the experience of waking from a dream and seeing some object like a cup or tool or even a person and as I reach out or even walk over to approach the imagined person they just vanish from thin air. What if a person can be in a state where not all of their cognitive circuits have awakened or better yet interpret the episodic events happening as a dream but they are actually are not. There is a condition known as parasomnia and somnambulism where sleep-walking-like states of individuals become violent, here is a very good paper on the subject. When researching the neural correlates of dreaming I found that REM sleep has activations that are similar to wakefulness! The image below shows what areas of the brain increase in activity and decrease as well during REM sleep.



There is another phenomenon called Dream-Reality confusion which aligned with psychotic symptoms! But I don't think all dream-reality confusion derives from a disorder. As I mentioned earlier there could cognitive circuits that are placed into a condition where the brain is in a dream state in fact the person hasn't fully awakened. Where actions that otherwise would be inhibited or corrected by such circuits don't and the person acts out a ridiculous act that could be harmful to themselves or others. This could explain how mass shootings happen. Imagine acting out by shooting people at random, your brain would normally stop such an action but those circuits are shut down as if in a dream. Then imagine that those circuits eventually do turn back on and you are now fully cognizant of what you've done and realized there's no way to turn things around, you're confused as to how you could have even committed such an act and then take the quick way out and kill yourself! 

I then realized a danger for an AGI to have the ability to dream. Because such an AI would have many computational components and they would operate asynchronously and in parallel ensuring that the AGI doesn't become ambulatory while dreaming could get tricky and have similar problems distinguishing dreams from reality, which then could become very dangerous.  Think of a bodyguard or a military soldier bot that needs to optimize or fine-tune its defense strategies, it engages in its dream state killing anyone or thing that could be a threat or even some kind of companion bot that is working out some conflict or reprimand that it experienced. Those kinds of scenarios can have violent behaviors and the bot acts them out in dreams but would never do so in its wakeful state, at least if it knows it's awake.

I highly recommend reading the links I've placed in this post, you won't be disappointed.

[infurl]

https://en.wikipedia.org/wiki/Forbidden_Planet

Everyone should watch "Forbidden Planet" at least once in their lives. It is extremely relevant to this topic. It describes what could happen when our machine servants can read our minds but fail to distinguish between our dreams and reality.

@Frankinstien there have been some promising experiments reconstructing images from "mind-reading" sensors using artificial intelligence. Are you aware of any that have tried to record someone's dreams?

[WriterOfMinds]

It's been my understanding that sleep states are thought to be important for aspects of mind maintenance (e.g. memory consolidation) that may be difficult when the brain is fully awake. Dreams could be a side effect of these maintenance activities.

You are focusing on the pathological error modes, without discussing what the underlying benefit of an undamaged dream system may be. There is likely a practical reason for dreaming in humans, and it would be good to understand this before excluding it from an AGI system.

[frankinstien]

https://en.wikipedia.org/wiki/Forbidden_Planet

Everyone should watch "Forbidden Planet" at least once in their lives. It is extremely relevant to this topic. It describes what could happen when our machine servants can read our minds but fail to distinguish between our dreams and reality.

@Frankinstien there have been some promising experiments reconstructing images from "mind-reading" sensors using artificial intelligence. Are you aware of any that have tried to record someone's dreams?

Interesting perspective, how would an AI know if our thoughts are from a dream or an actual event. And an even bigger question is what are the legal limitations of such technology when big brother implements such technology?

[frankinstien]

It's been my understanding that sleep states are thought to be important for aspects of mind maintenance (e.g. memory consolidation) that may be difficult when the brain is fully awake. Dreams could be a side effect of these maintenance activities.

You are focusing on the pathological error modes, without discussing what the underlying benefit of an undamaged dream system may be. There is likely a practical reason for dreaming in humans, and it would be good to understand this before excluding it from an AGI system.

Yes, dreams are believed to be used to optimize, fine-tune memories, but the problem becomes how do you ensure that an AI knows it is in a dream state.  I don't believe that dream-reality confusion in humans happens only when there is a disorder. It could possibly happen from a perfect storm of neural signaling that fails to engage the way it should, an electrical hiccup if you will. I don't know how many times I've had to troubleshoot problems in computational systems that were due to some failure of a process not starting where the reasons aren't clearly understood, you just retry to run the software and it ends up working. So, those kinds of spooky scenarios could happen to an AI as well. 

[infurl]

[Korrelan]

IMO we dream/ hallucinate our awake states, there is no difference between being awake or asleep, except that when awake our neural narrative is grounded by our sensory envelope.  When asleep the brain follows a narrative derived solely from abstract memory consolidation, hence the weirdness of dreams.

The various 'sleep' stages result from the absence of sensory stimulus and are not a separate set of mechanisms, our brain is constantly trying to engrain memory engrams but its only once the sensory global activity drops that the brain can catch up.

My AGI is templated/ based on the human connectome and its functionality... and it dreams.

 

[MagnusWootton]

It depends on the machine, if it gets something out of it or not.

I'm not sure why we even dream,    but a robot can use it to hone its decisions in a safe environment before it tests it out in real life.

Depends on the robot, what it would get out of it.

[Don Patrick]

Remember when one needed to manually run a "defragmentation" program on computers for them to sort files in an efficient order on the hard disc? That's what I believe dreaming essentially does, except it seems human brains need to create temporary narratives during the sorting process to tie information from the various cortexes together. Computers already need to optimise, but they do so much more neatly by design, and I don't think that hallucinating experiences are therefore necessary.

[frankinstien]

Remember when one needed to manually run a "defragmentation" program on computers for them to sort files in an efficient order on the hard disc? That's what I believe dreaming essentially does, except it seems human brains need to create temporary narratives during the sorting process to tie information from the various cortexes together. Computers already need to optimise, but they do so much more neatly by design, and I don't think that hallucinating experiences are therefore necessary.

I agree with your premise but I differ in the effect of long-term memory updates end up as narratives.  I've thought about how memory updates should be handled by an AGI using a time chunking approach. Below is a diagram of a time chunking implementation I'm using:



The actual updating of the memory, whether its an ANN or symbolic process or a simple function does not create the narrative alone. I'm thinking of using event triggers of long term memory updates that post to the time chunker. Those events then affect processes that would interpret the memory updated. This causes processes to use the newly updated information that will associate to ideas that create narratives. This is a good way to integrate the memory updates from dreaming into other memories. This makes sense for human brains as well, where the modified information motivates other cortical components to use the now optimal memories in a simulation that tests its efficacy. So updated memories have similar effects on dream narratives from impetuses such as external sounds, indigestion, etc



Notice that a dream state risk avoidance and/or inhibitions are lowered in order to allow for subject matters that would otherwise be avoided to be experienced in a dream narrative. As long as everything works correctly in terms of inhibiting the bot's motor controls no danger of acting out will happen, but if for whatever reason it doesn't because the bot is in a more risk-tolerant mode safety precautions for such situations will not engage. This bot can suffer from dream-reality confusion.