How will Artificial Intelligence Effect our Future?
By Diadon Acs
It seems artificial intelligence technology is inevitable for the human race's future. Despite what may be portrayed in many movies, A.I. has the potential to partner with its parents to create an incredibly abundant future. As more advanced A.I. succeed further on the Turing Test it’s only a matter of time before we birth a complete intelligence with complete self-awareness.
The question that many are asking themselves is one of social interaction and the exact role A.I. technology will play in our future. Will it be a detriment to job security, or perhaps serve as our gateway to living forever? These questions I can't answer, but I can give my opinion on what is possible. It seems very likely that the importance of what we teach our A.I. creations will be the greatest determining factor in our future. Advanced robotics with A.I. integration can help reduce workloads for large corporations, but what does this disruptive technology mean for the working class?
I will briefly discuss this subject as it falls more under the category of social engineering rather than computer/robotic engineering. It is an important subject to consider, however, since the two forms of engineering are now beginning to directly affect one another. There are two general propositions for seamless integration when robots out work humans and put people out of a job as they have begun to do. The first is a universal wage, in which all people have a set amount they will be paid monthly for all of their basic necessities (ie Food, Shelter, Water, etc)
The second is to completely abolish the monetary system altogether and create a gift based society, or something that resembles the venus project. There are probably several more options that escape my ability to comprehend. There is a 3rd option that saddens me tremendously, and that resembles something a little closer to a dystopian sci-fi movie. I will come back to the importance of these social engineering considerations later.
What are some of the problems A.I. is solving now?
-Cited from the business insider
AI is becoming more advanced and out competing with humans in many tasks. Probably the greatest example of this is Deep Blue which is a computer that beat chess master Garry Kasparov. This brings us to our first type of AI
Type I AI: Reactive machines
This type of machine learning is based on reactions in real time. With little or no memory or past experience to go off of. It uses real-time probability to carry out an action. It has been pre-programmed the outcomes of the actions taken and then reacts to the probability of the tasks that are possible. The problem with this type of AI is one of redundancies and specializations that don’t allow the AI to do different tasks besides the specific ones that have been programmed. These Reactive A.I. Machines are usually the most commonly used today for specialized tasks and reliability. A solution to redundancy is machines that learn from past experience, which bring us to our next AI Type.
Type II AI: Limited memory
This type of AI learning is pre-programmed with set parameters like type one machines. Over time it alters its programming to experience.
An example would be an autonomous vehicle which has been pre-programmed to drive in certain lanes and go to a set destination. However, along with its path from point A to B, it learns along the way with sensor data to ensure it doesn’t run into things.
The problem with this type of AI is that memory is limited in these mechanical systems. They don’t have the storage capacity like the biomechanical brain (ie Human Brain). So the stored memories of past experience are only shortly stored to conserve memory. These shortcomings are slowly being improved upon as technology advances. Combining type 1 and type 2 neural networks is a new step forward in the evolution of AI technology. Creating the ability to have two neural networks communicating with each other to critique and modify results.
Which brings us to type 3 A.I. using the theory of the minds network and simulating that in AI systems.
Type III AI: Theory of mind
This is a very advanced robot system that can connect itself to other paths of thought. Connecting to neural networks that already exist such as a parent passing down information to a child. Its called mind theory in psychology, and it is a cognitive ability of a network to recognize that other networks are present. Or in other words, that there are other creatures that are alive too. This is a major challenge for AI development to get a machine to recognize other entities without out being corrupted by there data. A good example of how this is still a major problem is looking at the results from Microsoft’s Tay Twitter Chat Bot.
So a major problem is training morality and ethics into an AI’s environmental interactions. I am sure as time goes on, there will be new ways to approach this problem using a combination of Type 1 and 2 A.I. Much like what we humans like to call a conscience, using feedback neural networking can potentially do this same process.
These types of steps for creating multilevel neural networks, to make a more complete AI mind, most certainly are the direction we are headed. It's just a matter of time and resource before we are able to start discovering the full potential of our A.I. children.
Type IV AI: Self-awareness
This is truly the end goal for A.I. and truthfully a spiritual journey most of us humans are trying to take. A self-aware entity is the essence of complete cognition and all life itself. So if this goal can be achieved, A.I. will be a new life form on this planet that perhaps has the potential to surpass our intellectual abilities.
An empathetic relation with other neural systems or beings around is crucial to the success of this type 4 A.I. System. For example, when a person yells “Watch out”, other people around recognize that there is urgency in the voice and perhaps danger. Most people can then relate to that past experience of urgency to put a greater importance that in a particular situation. This type of interconnection with the environment, as well as self-awareness in a computer minds place in a said environment, will be crucial to making the advancement of A.I.
Though code can create more code through templates and duplication, Those templates have to be written by a person in the first place. It seems almost impossible for a “new” template to be written by A.I. alone. The human aspect of creativity or thinking in new ways is a problem not yet solved for A.I. computing technology.
The greatest problem is probably Social Engineering
When this type of A.I. is achieved, the redundant specialized tasks of Type 1 and 2 will be easily trained in or out… but a new sense of morality or computer enslavement might arise with the advent of machine consciousness.
This brings us full circle to the original moral dilemma established at the beginning of this article. That is what will our social engineering look like when we are able to bring such advanced machines online? What are we planning to teach these creations of human engineering?
These are the really important questions we should be asking ourselves now. To avoid future crisis events that can arise from ignorance in social engineered designs. It's important to use natures designs that tend to flourish efficiently to our advantage harmonically so to speak. If it is nature we are attempting to emulate by developing A.I. Shouldn't it be nature we ask the questions in order to find answers for the progressive evolution of A.I.?
The development of Artificial Intelligence Technology lays in the hands of its engineers like any progressive technology. If you are interested in learning more with hands-on experience, here is a free neural network program I have enjoyed.
If you enjoyed this article please consider looking through my e-commerce store and check out what I have to offer. I have some amazing deals and unique apparel. I appreciate your support and encourage your thoughts on Artificial Intelligence Technology. Feel free to leave a comment down below.