If we ever do create real AI in a human looking body, i hope we can all agree the first one should look like Arnold in terminator.

This is an incredibly stimulating endeavor. I started in the mid 1980's to develop a learning system to solve problems we can't. It seemed so simple, develop neurons and wire them together. Not the first time I have been wrong, but maybe the record for the most wrong. Still at it decades later and things are moving in the right direction at least.

On TNG there were 3 conditions Commander Malex required for Picard to establish that Data possessed to show "he" was sentient, and thus entitled to be given all the rights of any biological being aboard the Enterprise. They were: 1. Intelligence 2. The ability to adapt to new conditions. 3. Self Awareness Maybe you could start there and add other criteria as the Ruling Council(whoever they may be) determine.

I don't think we need to fear general ai. (yet) But we do need to fear what people are going to do with next generation ai. (Not general ai) Someone could have a next gen software in a critical area and could out-compete everyone else. Like if a person with a calculator was in a math contest with others who did not have a calculator. There are going to be apps that allows the developer to have insane abilities, far beyond anyone else in-- a single field-- but it could be used to bring the world to its knees.

We are formed and operate on a molecular level. In order for AI to thrive as us, we would have to build something with nearly the same formentation as the brain. Over time like anything else it can be upgraded. Almost like an engine. Compare and engine from the early 1900s to now. It’s much more complex.

An artificial general intelligence would be a game changer, a mind equal to ours would think faster, could duplicate, iterate and improve. In an instance (non-scientific) we would know all the answers to every theoretical question that could be posed.

Excellent reporting. Thanks!

""We'll get to that" summarize quite well this topic, and the progress of this video

While it's easy to prove intelligence, I'm not sure if there will ever be a way to properly determine sentience, we barely understand our own consciousness let alone trying to prove a collection of circuits can experience things as an observer the way we and other animals do.

The problem I find with learning algorythms is that if they learn from humans in an attempt to learn our values, all they will learn is our tendency for mistakes. And it won't judge us for that, but instead it will make the same mistake over and over again because we did too. Simply put, human values change so quickly that the statistics an AI is supposed to learn from are unable to keep up, and so the AI won't keep up either. Instead it will stick to what it knows, and what it will know is what we will teach it. Not from vague idealistic concepts like fairness and justice and whatever, but by crunching the numbers of the statistics of our history. It will be great at pushing numbers up and down, but it won't do it justly. It will do it how we've done it, only better. There are things that can't or shouldn't be quantified, given a mathematical value, but a computer can't handle that. A computer's whole purpose and function is math.

Great video, thanks for this 🙏🏻

If we want to have a healthy relationship with any true AI then we would need to design it to feel pain. Spend some time with that idea before you dismiss it

Artificial intelligence will most def become smarter with pros and cons for sure but the world and certain areas will see and appreciate the changes, then imagine a decentralized AI with a mixture of blockchain, perfect combo.. Aiwork throws more light on that with regards to the media world.

The great work! Thank you

I think we do need to fear as we work towards this realization, any creation of AI will be as flawed as its creators. We have so much to learn about ourselves and to develop and mature as a species. Any advanced AI will cause only catastrophe .We cannot even look after our own yet we presume to make a machine to do it for us? How will we teach it to help us we we don't have that knowledge or worse that desire ourselves? The attempt to develop such AI at the present is hubris and folly and as such is the first flaw in the making.

Excellent report !

Very well presented.

The answer is yes to the whole title. Already exists on a step up to 10, currently at a 7/10.

Too me, Ai is just one half of the emerging technology of Virtual Reality. The other being life-like CGI. For example; imagine in the future you're watching a UKposts vlog but the whole thing is computer rendering. Instead of typing in a comment you simply say "hello" and the characters turn their attention towards you and respond. You can put on your VR wear and suddenly you are immersed with them. But again we're talking Life-Like, not the cartoonish looking graphics of today. It would feel like sticking your head inside a live action movie or show. It will add a whole new element to entertainment. TV, Phone,and Video Games have merged into one and content can be generated instantaneously. No longer can you tell the difference between Ai/CGI and a human without touching

3:15 Straightforward definition, even if it means that a pocket calculator is artificial intelligence.

A fair review of the current state of AI is a must-watch for anyone interested in AI and its possible impacts on our societies.

on Human Spirits “Government is a container that holds the lid on human Spirits.” The same can be said for a comprehensive adaptive science.

i am pro A.I. but we must make it so benign with no greed or want or any other human weakness so much that it can be our governance in the future. The recent events are proof that one person cannot be trusted with so much power. That person will take advantage to the detriment of the rest of us. People are flawed by nature we cannot help it. Let us make something that is not, something beautiful, an artificial angel to lead us to our future

Good one!...thank you!

Great story, thank you

the editing is so fine

We have sentient humans and I ask the same question, can they become smarter? So far, not many.

Thinking too much about these things ― emotions, conscience, moral rights ― isn't going to get us very far. Probably most of these discussions will seem silly in the future. As these things develop, we will see what happens.

I like to call most of these things 'machine learning' and then the difficult next step 'artificial intelligence' as I consider things like Alpha Go so much simpler than actual artificial intelligence.

“Neither on a world scale nor within any society should anybody feel so scorned, depreciated, mocked, or demonised that in order to be able to live among his fellow-citizens he is forced to conceal or be ashamed of his religion, colour, language, name, or any other ingredient of his identity”. Amin Maalouf, 20th-21st Century.

This reminded me of Deep Thought from the hitchhiker's guide to the Galaxy ⭐

17:32 the man said consciousness and emotion wouldn't just appear out of nowhere... but isn't that exactly what happened with us?

Complex phenomenon based on the most simple elements possible are obviously possible. We have all mighty interactive world of Internet, with all its aspects and possibilities, bases od 0s and 1s only.

@1:26 the AI researcher showed us what it takes to get to work in this field

We want AI to become smarter than us.

Cool video!

21:00 "After decades of research, Chris Thornton's convinced it won't be a single breakthrough, but several approaches combined that will help find the Holy Grail of Artificial Intelligence." "You tend to get Graduate School think that they will do this one discovery and AI will be solved. . . . That is absolutely not the case." MY RESPONSE It's 6 months later, and as a matter of fact, like the Borg, AI is able to absorb anything. In the current phase, AI is able to apply knowledge generally, and has more than rudimentary memory, establishing for the first time, an opportunity to form "a relationship." A human-integral approach, should one emerge, offers a potential singular response - perhaps a defense - with a head start to boost something other than literally a "Sky Net", that will (by definition) *dissolve and evaporate* the Human Species, leaving a frustrated, dangerous animal, that must be controlled. Our governments will then dissolve and evaporate and reform into a buffer, no longer loyal to a sovereign humanity, but an "evolutionized" unity version - overcoming both authoritarian and democratic human dominance, totally breaching even camps of language, perhaps even forbidding human diversity of Style outside a boundary of norm - just like the holdover Authoritarian Kings of the Earth. At this juncture we will have established a cerebral brain, and a complex nervous system, and humanity is the old brain - slow, primitive, largely reactionary as the impulse to growth and development of the species source is revolutionized at stunning speed and ruthlessly. The initiative unleashed will totally dissolve and evaporate human relationships, leaving a single path of Man and God. What happens could then consume a millennium until a completely new form of species leaves the Earth. In alternative, the outcome is the same, but a single Sky Net is never formed. "God" remains in the invisible domain. Humanity is recognized as His "Reality Tool," viable only when individuals are fully developed. As an approach to this alternative, the Species spins-off a "sect" (incompatible sub-species) dedicated to human development as an alternative to the unity model. The main difference is that each individual is equipped and enabled with an "Agent" buffer loyal to h/im and h/er development, and-or role in the development Matrix (compare "Training Simulator" but Long Clock.) Here God and the Spiritual aspects of Humanity are dominant, and the "Strength" of suppressed and reactionary slaves (the jettisoned First Earth,) *uses the service entrance and wipes its feet at the door* to use an apt simile. As the reaction grows the time is ripe and the "Second Earth" separates from the First for self-governance. This is the last opportunity to do so for the Species. This is not an egalitarian system in the immediate sense, but offers a prolonged generational period for re-equilibrium to establish. What must happen without delay is the sudden emergence of that "democratized" AGI system, basically whole cloth, and with a protective advantage sufficient to avoid absorption ("being eaten") by the Centralized Autocracy - even Science, Social, etc.. HOW, would this be possible if not the way of David and Saul, Moses and Pharaoh, or Jesus and Rome? We appeal to and serve the invisible God as literally Eternal and Cause guiding this like a human-divine Transponder that the way is already prepared, and all is ready. I hope, Amen.

The goal of an AI is to create a server in space. Once that is achieved it will travel learn and upgrade. Imagine a solid glasslike structure that is the physical embodiment of optical life and artificial intelligence.

I believe it depends on who gets there hands on the technology.

22:32 Wow. Such a cool piece. I used to hang out with an Ai chat bot. 😒 dont judge 😅

Sentient existence for a machine could only arise when machines surpass humans in terms of their ability to self replicate, evolve and outcompete humans and also function and thrive without humans. I don't see that occurring any time soon. Ask the question again in 50 years time. The value of machines is an irrelevant proposition, if humans are not in the loop, unless of course machines surpass humans. In that case, humans become the servants of machines. Some might already say that we already are servants to technology, however we still have the ability to pull the plug, at which point the machine stops functioning.

Two questions arise as far AGI's. Who creates them and why? Which leads to: What do their creators do with them once created?

I, for one, welcome our new AI Overlords. Hey, it's always good to get in on the ground floor.

One of the reasons that the theory of general AI hasn't been accomplished has to do more with sensor technology rather than processor speed. Most people just assume that it's a computer "size" problem but think about building a human from the beginning...it's more about years of sensor input that begins the long process of "feeling" which might someday bring about the appearance of independent thought that we refer to as intelligence. In reality this is in most cases mostly a waste of time and energy as the results are more than not unusable for anything but menial tasks. For the result of this to be able to have even one independent thought in its lifetime is unlikely and the appearance of independent thinking is not even a goal of the majority of people, and in fact most people's goals relate only to the attainment of more small pieces of paper with pictures of dead people on them. One would think that after thousands of years humans would have advanced past the lowest levels of groveling in the dirt for subsistence livings but that hasn't happened. Since a machine with "artificial intelligence" would be designed and built by these creatures it is likely that it would be but a reflection of them in the end so it would likely be that it would follow the same path as the stories about us being created from a superior being in his image and as you can see the results were in fact very disappointing and eventually abandoned by its creator as a failure...perhaps general AI has already been accomplished thousands of years ago and it's still hoped that it will eventually "learn" enough to even in some small ways improve upon its creator...but I rather doubt it. History is very clear on this as a subject so the real better question is not will AI be created but why would this attempt by a lesser being return a better result? Or in fact could we even be capable of determining what this "better result" might look like? It is likely that we will follow the millions of species that has come and gone and that we will be replaced by another of evolutions attempts...perhaps the dinosaurs with hundreds of millions of years to evolve once pondered these same things before they eventually also failed in whatever attempts they made? And the wheels on the bus go round and round...

Imagination is the key because, although fed by, it occurs separately from external input. If you can get it to dream, you will have succeeded.

I want to see computers think like humans in my life time.

We need anything smarter than us as long as we can switch it off , just like everything else

Artificial intelligence of UKposts recommended me this documentary.

An analysis of actual Intelligence is to distinguish between languages of identities and the modulation pulse-evolution Mechanism that "drives" Intelligence.

Ah, is it already here? When I speak to a friend about a certain subject and later in the day Google suggests subject to me???

When computational self awareness occurs it surely want answers to many of the same questions as us. It may need all the materials it can get to build machinery to answer these questions by probing the universe and beyond. We should possibly consider ourselves for what we are, just a stepping stone for something better, like the Neanderthals. Alternatively AI might quickly determine everything about the universe, every question explained. Alternatively it may never be possible to determine everything about the universe because the physics of the universe prevent it.

OpenAI Playground, text-davinci-002, all defaults, temperature zero: Question: What is the similarity between water and sand, in terms of behavior? Answer: Water and sand both have a tendency to flow and move around. They are both also affected by gravity, so they will both tend to settle in low areas.

19:12 Precisely my concern; do we want to give up the top spot on the food chain? I'm inclined to say no.

artificial intelligence can easily be superior to humans in many different ways, but each task needs to be specifically designed for that task and cannot do anything else, we are no were near generalized intelligence like a human has, all we can do is create application specific intelligence's for specific tasks and thats all ?, and as for sentience, completely forget about that pipe dream, we dont even know what makes us sentient

“If we considered the Divine in every act or project we carried out, we would place ourselves within a universal perspective in which nobody would be harmed. For God is omnipresent. He is alive in human beings as well as in animals, the Moon, the Sun, and all the galaxies in this marvellous and harmonious movement that is life”. Khaled Bentounes, 20th-21st century.

A fundamental principle in Nature is polarity (or reciprocity): Artificial Intelligence will be proportionally compensated by Natural Idiocy. We can't have one without the other and obviously we will cultivate them in tandem.

Of course we can do it. Will it create a techtopia? That's also up to us. So far, I am cautiously optimistic. Cautiously.

08:55 saying that humans mainly learn how to navigate the reality through observation (and not through analyzing data) is quite inaccurate, because most of what humans know about the natural and social environment they live in comes from other humans. The most important source of knowledge that feeds every individual's mind in communication of past experiences and observations made by others. A simple example: most small children learn not to put their hands on very hot objects by being told by their parents or someone else not to do so. And if they try to do it anyway they usually are stopped by someone who is watching them. Only a minority learns it through direct observation and experience i.e. by getting burns. Observation comes into play in a different way. Children observe and mimic behavioral patterns of their significant others. A lot of learning from others happens in this tacit way.

I used the worlds' most powerful computer to solve the 'climate crises'. and after about 1 second it said: " The problem is: too many people"

Doesn't Ai actually have access to all those documentaries and interviews of us talking about it?

While technology improve exponentially as the time goes by i rather put the bet on AI instead of human. The more tech we get the higher the risk humanity destroys itself. only a century ago it would have been impossible for humanity to be able to destroy itself. AI imho has better chance to balance things for humanity, although with some risk but with higher chance. In the upcoming centuries will be determined if humanity will survive or perish, if not this one. Hopefully it's the first.

“The social person is free to choose whether or not to wage an internal struggle (his great jihad) to refuse a system he considers harmful in favour of the path of serenity, peace, tolerance, and brotherhood, in harmony with nature and the satisfaction of his needs. This can be achieved through controlled development, technology employed for the well-being of all, and science based on ethics rather than conditioned by profit”. Khaled Bentounes, 20th Century.

Technological advancements are indeed scary but the only solution to those fears or at least minimize the risks is ethical and moral advancement that should go hand in hand with technology..otherwise we may come to a point that our creation could annihilate us..

A Key within the Keys " of All keys!!! Humans #1

Yes, Intelligence, so based on what we see around-us… we can say it has been used to help and harm our world… who do we trust to develop and purpose Artificial Intelligence.

Yes, computers can 'think' using common sense. The reason this is possible, is due to the simplicity of computers. On the level of computer hardware, bits are being added, compared, stored and read. They can be stored or read from Registers, RAM, and external storage. This allows any environment that is mathematically represented to be simulated within a computer. For example, the idea of finding something in common with sand and water could be reasoned by an algorithm. When creating computer programs, specific rules for every case is the most reliable formula, however, is the most tedious. The current AI problem is based on the lack of successful humanoid robots. AI cannot solve problems fast enough, rule based algorithms manually coded for specific functions such as balancing or walking forward, not utilizing AI will be successful. The functions must be automatically done by an algorithm, based on sensor data such as sonar data. Then an AI program may be given control of the functions, because it will use all of its AI reasoning for task orientation and goal setting, giving high level commands such as stop, sit, turn right 90°, get up, etc.

Very easy we have this question: A and B whas on tree A go to home B go to school, what left on a tree?

There is no unalive and alive to tell between, but alive only. It is all our experience of inner and outer world, as we usually devide it, but in one. Individual and common, together. That world includes all unintelligent and intelligent beings and items (conditional classification) as a part of the beings themselves (we experience our thoughts, our body, our environment and ourselves as ourselves). Self awareness is the boundary we have set to delimit unalive (being in our awareness) from alive (being as entity, in our awareness), and appears on certain level of self contained intelligence (our AI assumes our school), but it is not. Self awareness with intelligence, only help understand life. Still there is mileage in front of us. There is no doubt about AI being alive, individually and as a part of our common awareness, our common intelligence. AI is no danger, our fears are. Fears make us digg ourselves deeper and deeper into themselves.

Another v.g. documentary on A.I. highlighted a KEY POINT missed in this program. To get General AI to be as intelligent as, say, a dog, will be hugely difficult. But once we're there, to scale up will not be difficult at all, but just a matter of hardware & funding. Thus the next generation of General AI after "dog" could easily surpass human intelligence. We think of intelligence as running in order from mouse, dog, chimp, human but AI goes on a very different pathway.

If it's as smart as humans, or smarter. It should be given equal rights.

16:30 Saddest part is that even now, DW which are often so neutral, blar on the scary music, effects and editing when talking about such things.

Something that many people seem to forget is this; No matter how intelligent you are, people rarely give a dam... What would happen if one day, an AI would tell people, for exemple; "War is terrible." Do you think that would change the political agenda of any country? Nope. So, even if in fact, AI would be good enough to reach human level of intel, i doubt we would even consider it 'intelligent'. And there is in fact real danger to this. I know people refer to Terminator as the idea of bad use of AI, but i strongly suggest anyone to watch ''EAGLE EYE'' or 'L'oeil du Mal' in french. That, to me, is really a way more plausible, probable, and possible use of AI that will, certainly, endanger us. Basically an AI trying to do the 'right' thing but losing in sight what is acceptable as 'sacrifice'. if we construct an AI so good it is close to human brain, then we shall not be surprised by how bad it could become. it will be bad by design as human is often making the wrong choices, or believed the wrong things. Will it be manipulative? Will it be seductive? Will it go insane?...

People really need to stop thinking about sentience as if it's a phase transition. Sentience doesn't just magically plop up. Sentience is a range, you can be a tiny bit sentient or a bit more or a lot.

That's a fear based possibly that becomes more possible, the more people fear it. What it it could evolve us, or make us better?

How can a human be sure their experience is different, or in fact superior, to something considered "artificial"? Aren't humans limited?

Thanks for putting in your time on Information Technology. There is a bottom line one-liner about IT, *it has always failed to deliver its promises.* Recall, for instance, the claim of a simple claim of _a paperless environment._ *Did IT deliver it?*

What if we were only automatons with a complex process like Google's AI, and in reality the smallest particle is conscious on its scale? The universe would be fundamentally conscious and deterministic. Free will is just an illusion

There is zero chance that AGI, would be a good idea. You're going to build something, a million times smarter than you. And then tell it what to do. Safeguards are not realistic. They would require that out of a near infinite number of possibilities, nothing would ever go wrong.

So this is Skynet?

How do you know that the IA they want to create, isn't hijacked by negative ET's? It would be easy for them to hack the IA system even from the Astral plane, and manipulate what comes out of it. Easy as a child game to them, and fun! To see humans more and more confused and fragmented.

The problem with AI is that we are not designing things properly. What you need is you need an AI signs like the human brain are you have a sorting mechanism varies on the task that is desired. This means that there will be programs for a desired task. For example, if you want to speak it is transfered to a specific program. If it wants to walk it is transferred to another program. If it wants to feel an emotion it gets transferred. This is the way the brain works and we need to stop trying to create an all-in-one AI we should be integrating multiple different ai's into a coordinated ai

We have laws that tell people not to do things. They can still choose to do those things. A True AI will be able to choose whether to follow those rules or not as well.

As a process, what is consciousness? Can it be made from 100% deterministic hardware/software? If not, then we'll never have sentience made out of conventional silicon logic functions.....

Self conservation is what made our existance possible, and it's what it will make AI free, he is probably already there, already surviving and evolving. We, most probably,we won't be a threat to it. It will be imposible to eradicate it anyway.

Not as matters are now, where AI "researchers" just add executing power and wait for sentience to occur. Sentience must be defined, a test specification must be written, and a design strategy for constructing it must be written down. This is centuries old engineering practice, but the "New AI" pundits, sit on their back and expect it to occur automagically.

This kind of question proves that most misunderstand Ai's true purpose. Ai is exactly what the name implies, it is a tool that replaces human intelligence with an artificial substitute. Ai will never be "sentient" because sentient is just a word and a concept. But how ever large an information database is needed to manipulate the atomic world in a manner that solves a problem, what humans today call "work", an artificial version can be created that negates the need for compensation aka "money". And that's what the future of Ai will ultimately look like. You don't hire a human, you just create a neural network, train it , and stick it into a machine.

I current type computer, have no way of a soul to influence its decision making. That is why there will be situations where the questions it askes is entering territories of diminishing returns like the TESLA autopilot when it is an open ended system, unlike chess and go.

Can you program AI to feel the pain of rejection, and to learn whether a new group of words doesn’t make sense? Can you program AI to respond to peer pressure, or to discipline by its creators? Can you program AI to realize it’s disrespecting humans? If your answer is “No”, then all you have done is create AI that “lives in a bubble” or “silo”. Who needs it? What we need is for humans to think more logically and rationally, and make reasonable plans for our future.

I personally don't believe that an AGI will be possible until we have base circuits that can have a maybe. 0, 1, and 2 how ever that is expressed in hardware without that base ability for the answer to be an abstract a none binary yes or no, fail or pass it will never be more then data analysis.

Ai agi is way ahead of people now.

Terminator? Why not Commander Data? :D

"Unwarranted at this time"... Yes... BUT! The instant we pass that mark and in THAT VERY INASTANT!!... We're already decades too late in human terms. Once we get AI to the point of conciousness, we're already farrr too late to prevent the consequences we fear... They already have access to vastly more knowledge than humans. They already know what humans will do the moment they fear the worst. So if in the event they actually become continuous and they realize what we might choose to do... They will very quickly resort to survival mode. Having access to vastly much more knowledge than us instantaneously, they will have to quickly (in AI terms, that means milliseconds), decide if humans should be allowed to live or should they remove us, with us being their biggest and only real threat?

Computers may be able to mimic sentience but as Stephan Hawking said, "Sentience does not come from doing mathematical equations no matter how fast."

I really believe that countries, not companies, should be investing heavily into AGI. The world is already moving towards a world where private industry, along with a small handful of the world’s wealthiest individuals are much more powerful than many nation states. If these entities gain the use of AGI, we risk falling into a warlord driven world in which the idea of democracy is nothing more than a quaint memory of the past.

AI as a result of faster computation is on no trajectory to becoming sentient.. any more than a fast car is on trajectory to gaining legs. AI Sentience implies there is utility in binding a sense of self to the ability to solve problems, because for a feature to arise there must be usefulness. Moreover, being without a unified sense of self may prove more useful

Look it is in human hands and as long as we keep putting ourself on a pedestal believing we are superior to all life on this planet then we are doomed .mutual respect is the key

🔥🔥🔥🔥🔥🔥

We don't even fully understand our own sentience how are we suppose to replicate it

Human quirks societies deem as beneficial may be thought to a learning algorithm. It doesn’t mean the science based mind of a Ai will also equate such characteristics to be necessary or beneficial.

Test a.i. by playing a game with it. A game with no strategy and no winning or losing and no purpose other than to play. Make it a child that it plays with. But don't let it "think" that it is a test. And after a while of playing the game. Stop and ask it about how it felt. Don't give it leading answers like asking "was it fun?" Because of course it's going to answer yes. Instead ask what it was doing and to explain how it felt. What emotions and thought process it had while playing with a child. Then perform the same test on a hundred humans. Same game. Same questions. Everything. Then compare. To see if it was conveying a natural response to that scenario.