Deep Learning AI Taking On Human Creativity

I talk a lot about artificial intelligence, it's potential, it's promise and it's danger these days.  Most individuals I speak to agree that AI will allow for most manual tasks to be automated very soon.  They also agree that AI will be able to be infinitely more efficient at doing tasks like transportation, cleaning and construction.  With some explanation, they concede that AI will be highly efficient and better than humans at doing tasks requiring some creativity or intuition like guessing, analysis and speculation, because those can be trained using the large amount of data available through the Internet.

Then people have a very hard time even considering AI would be able to be artists.  As they say:  this is the realm of humans.  Only humans can be creative, they say.  Otherwise, what would we be doing in a number of years?  How would we be useful to society if the last type of activity where we excel at can be done better by artificial intelligence?

Of course, this is a reaction of fear.

The truth is, AI laboratories are already experimenting with deep learning AI and have achieved interesting early results:  AI can be trained to be creative and make music.

In the following video I discuss a bit more about the social and philosophical implications of this advancement:

What creative AI needs to get started is inspiration.  The sort of inspiration they need is similar to what human artists need really:  life experience.  In this case, since we're not interested in spending years teaching the AI about life in general, the scientists taught their creative AI about the type of art.  In essence, they gave them a huge amount of music to digest, from different styles and then through code and algorithms, gave them the ability to make completely original music.

What these AI are capable of is basically the same as what human artists are able to do:  get inspired by what they "know" and create something new.  They use algorithms called deep learning algorithm where the AI has access to large amounts of appropriate data for the task and then set to "learn" how to play with the data to get target results.  If the training period is successful, the AI is able to get the results we were interested in with the data on their own and can continue to learn over time.  Neat huh?

Here are a couple recent examples for you guys to check out.

First is the example of the Marimba-playing robot Shimon from the Georgia Institute of Technology.  It is fed four measures of music and then it goes on its own to create the rest of the music on it's own, inspired by both the first measures, which provide a specific desired style, and it's huge databanks of recorded music.  See it in action here:

Second is the example of Amper AI, which is a commercial venture that allows artists to create music by providing the AI with a style and tempo, and it'll create music on its own.  This is a beta product that has already been used by artist Taryn Southern to make new songs.  After the AI has produced the musical score, the person using Amper AI can adjust and layer on lyrics and other pieces of pop music to taste.  Here is more information about the effort.

Music made by Amper AI, for example when asked to make music styled after the Beatles (lyrics were provided by a human being):

Here is Taryn's song made using Amper IA, after made into a music video (lyrics and video by Taryn.  Instrumental elements are all Amper AI):

With great commercial ventures out and fantastic research on creative deep learning AI, we're sure to see more and more people using it and of course for the technology to become way more refined and advanced in a very short amount of time.

Stay tuned and for those interested, you can start playing with some of these projects yourself!

Super-Intelligent "Robots will outnumber Humans in 30 years"

Don't take my word for it.  This comes straight from Japan's SoftBank Group CEO, Masayoshi Son.

If you've been following along my blog and videos, that statement shouldn't be very surprising.  For those of you just tuning in or reading these lines, here is the jist of it:

• Around 10 years from now, one of our computers will be able to do as many computations per second as a human brain.
• By then, there should be over 20 times the amount of smart devices as there are people on the planet, all interconnected (what we call the Internet of Things (IoT) nowadays).
• A few years later than that, there would be more devices, in more parts of our lives, and each one of those devices could potentially be as powerful as a human brain (though most would not need to be to assume their role).
• 29% of the 140 acquisitions of Artificial Intelligence companies in the world in the last 5 years were made in 2016 alone.  It is anticipated that there will be 300% more acquisitions than that in 2017.
• This year, China has already started using the first autonomous Security Robot, beyond the facier industrial robots, that can track bombs, criminals and other stuff as well as identify potential criminal intent on the fly.  I think it can give out directions too, like any good beat cop can.  It's the first robot that I know of that is in actual use in public.
• Autonomous self-driven cars are on the way by all major manufacturers along with outliers like Tesla Motors, NVidia,, Google, Intel and several others.  There are already self-driven vehicles on the road in England and other European countries as well as many self-driven trucks in the mining industry etc....  Major manufacturers for privately owned-vehicles still plan to start selling their personal self-driven vehicles on our around 2020.
• And of course you are aware of autonomous delivery systems being tested by several large companies (Amazon included) around the world.  UPS has joined the list of companies in this game.  A bunch of other entrants are giving a shot at the delivery game too, such as LastMileRobotics.  Of course, they likely would get acquired or become suppliers for larger firms if their designs hold up.

So with all that activity and integration of really smart devices and sensors running about... in 30 years, robots will be in fact everywhere in our lives, and they will likely be in larger numbers than human beings, and those robots will be severely outnumbered by smart devices embedded into most of our home devices and infrastructure.  Not everything smarter than us will actually look like a robot and vice versa.

I talk about how this will look like in the video below:

In a nutshell though, in 30 years, it is very reasonable to assume that we'll be surrounded with robots of varying intelligence, size, purpose, including possibly nano-robots in our bloodstream.

What does it mean for us simple dumb humans?

Well, it means we can enjoy life then because all that interaction between super-smart robots means they can basically take care of running everything we don't want to run, more efficiently, and with more competence than all the experts of the planet combined.

It means, sharing space with these robots and enjoying ourselves.  We no longer speak of work.  We speak of having occupations because believe it or not many humans want to be involved as a driving force behind our art, culture, science and society.  Humans will be basically fully integrated into a network of intelligence and robots.  

And don't think we won't have robotic lovers (mechanical or organic in nature).  Even today we've figured out how to give robots "senses".  It is thus quite reasonable to assume in 30 years, we'll have mechanical robots fully capable to feel and even be sensual just like a real human lover would.  

And why not?

New EU Laws Forecasting Self-Aware Artificial Intelligence

Stand aside Asimov, in the real world, we need real laws to manage advances in artificially intelligent robots.

And so we almost did.

The European Union Parliament has ruled this February 2017 on establishing some rules (not laws yet) to manage the upcoming robotics and artificial intelligence development in order to:

• Control liability related to self-driven cars and people transportation drones.
• Liability around other robots and artificially intelligent entities that may learn and behave in unexpected ways (due to deep learning algorithms and other methods).
• Help with the transition between a human workforce and an increasingly AI-established workforce.

I've been talking about this for a while now and I'm quite happy to see that governments, who are typically very slow in acting on technology, are actually taking decisions before AI enabled robots and software become truly able to learn and go beyond their own programming.

This is likely a reaction to the upcoming movement in the car industry (BMW, Ford, Chinese transportation drone companies, etc...) to put millions of self-driven cars and drones on the streets around the world on or about the year 2020 (3 years from now).  There is obvious fear about how people will react when the first self-driven car will occur or when the first AI enabled industrial machine hurts someone.

And what about the sex industry?  What happens when people panic because the economy can't sustain itself with too many efficient AI's doing a better job than any human can, taking their place in the world?

Lawmakers, as is their instinct are mostly thinking about the economics and the liabilities.

I talk about this and where we're heading with all this in the following video:

In a nutshell, lawmakers in the EU have rejected the Robot Tax  which is good.  Though it is a simple way to tax the work done by an AI robot that replaces a human in order to retrain said replaced person, but the industry would be stifled with such a measure.  

Instead, I am happy to see that the ruling is moving towards the installment of Basic Income for all instead, which is considered a riskier proposition but if we know AI robots can take most of our jobs, it means they can run the economy wheel instead of us, providing needed basic goods and services  to the people while citizens receive a basic amount of currency to trade for other things.

What seems to be a more complex discussion is on the accident liability aspect of things.  They are saying that liability would go more on the owner or manufacturer of the AI robot most of the time unless the robot has been running and learning on its own long enough to become essentially responsible of it's own actions, ergo being liable on its own.  So they're thinking of making regulations to enforce the use of "Kill buttons" on every such robots so that we can shut them down if they go astray and risk harming people.

Now this all brings us into sci-fi movie territory doesn't it?  (I, Robot, Terminator, etc.....)

In my opinion, if we're thinking this way, we have to start thinking of when an AI actually could be considered having rights of its own and is considered a conscious being.   If we consider a robot entirely liable for it's own actions because it has learned enough, it is just as liable as an adult is in the human world.  Before that, it looks like the robot's "parent" is liable for it's actions.

Very interesting and a concept I truly welcome myself.  There is no reason why we can't declare AI's as conscious and have them agree to rules that are beneficial to all.  

Thought you would not be alive when sci-fi became reality?

It's just the beginning!