Artificial intelligence & genetic algorithms

Even by now there's a lot of research going on on AI and genetic algorithms. Basically the tremendously increased computer power has already reached levels, where sophisticated computer software can learn and accomplish extremely complicated tasks by themselves. In 2018 there will be several applications, where computers won't be "running" pre-programmed programs anymore, but "learning" what to do and "evolving" programs for themselves on their own. The instructions given to the computers for doing so are called genetic algorithms.

Genetic algorithms are relatively simple pieces of software code, sort of "seed instructions", which just give the basic parametres to the computer on how to evolve. Once started, the algorithm will start evolving depending on the given instructions. If the given seed is chosen well, the algorithm will develop in a desired way, and the computer will learn the optimal ways of functioning. But even a tiniest change in the seed information may lead to completely different results - very easily to even catastrophic results! (Butterfly effect rules here...)

Genetic algorithms may be applied in forms of artificial intelligence, although also other types of AI do exist. However, genetic algorithms are often very unstable to the choice of seed information. Use of genetic algorithms will create completely new potential applications for using computer power. But they will also create potential for major misuse and catastrophs. Therefore giving genetic algorithms a role in the movie would enable major changes to the world of information as we know it.

Just consider a misbehaving network of supercomputers or a malicious genetic virus inserted to internet... You get it? The whole global information system might go crazy or collapse within minutes just due to a few (well) maligned bits! And that isn't even too unrealistic.