As we work towards the singularity, where Machine-consciousness will overtake the human one. So to help wrap our brains around this concept, we need to have a closer look at the areas of AI research and some applications. So just quick review and then looking at how to use it.
An Expert System is created (or coded) by a human or team of experts to complete a single task well. Well, because they are human created they can be exampled to predict the behavior of, but they are only good for the really simple stuff, as they can not deal with complexity.
Neural Networks simulate the neurons, but the structure they create is simpler than the human brain because of a lack of computing. They have a set of inputs and outputs they can react to an environment, and a number of hidden layers between the two ends. The more layers, the more complex problems it can handle. However, they also need huge amounts of training data to feedback, so It can get the right answer.
Machine Learning (ML) is closely related to Neural Networks where it can learn automatically based on experiences. There a re 3 ways machines can learn. Examining a massive data set top see what dropped out (aka Unsupervised learning), guided with a teacher directing toward a goal (aka Supervised learning), or reacting to a dynamic environment with positive & negative responses (aka Reinforcement learning).
Generative adversarial network (GAN) a subtype of ML. This is where a generator tries to produce responds that are accepted by the discriminator. One example is a generated photo of a person, which is part of this X does not exist.
Genetic Algorithms use massively parallel computing to evolve generations of algorithms towards the best answer. Each is tested and the bits of the best are mixed together for the next generation. These are great is they are multiple solution to a problem and have a clear goal. For example NASA used it to build a super efficient antenna, mirrors design, walking methods for machines, aerodynamic design.
Using Technology as a backdrop
The stuff described above can be used as techno-babble to make the stuff in your game feel more real and to help with ideas. The clearest example for this kind of technology is as an enhancer for people. Allowing you to do things faster. So least add some AI weirdness to your life by using this X does not exist to create some stuff for your game;
NPCs with a photo and resume
Snack food
Startup Companies
Fake News stories
Song lyrics
Save the page, if you wish to use it as it will be gone when you refresh or look away. Over the next 25 years (until the Red), these systems will only get more complex and create more useful stuff for use. Although there is something waiting in the wings…
The overlords step in from the shadows
Current AI systems can be considered a magic black box that does something, but most of what I describe here is Narrow or Weak AI. We set the goals for the AI and keep the results contained. However, the next step is Artificial general intelligence (AGI) is hypothetical for the moment. Although once we work out how to create it who know what will happen.
The first winner of the Squirrel AI Award, Regina Barzilay, “We’re not ready for AI”. The current pandemic is a great example of how we are not ready for the big events, let alone the Deadly Truth of General AI.
AGI, can think for itself and like everyone else you know they can work with you or against you, by setting their own goals. When dealing with children, you can out think them and stop them from taking a dangerous action. Humanity has a long history of raising children and influencing their decisions.
AI could be just another person in your game, but it could also be much more…
