Artificial intelligence is now used in our world non-stop. It is growing faster than ever, but few people even realise the rapid rate at which AI is growing. Eventually it will reach the singularity, the point where future laws and rules are not understandable by the generation prior. This is the point of no return for AI. This is predicted to happen as soon as 2049.
Google is the most well known company to use artificial intelligence. Often they are at the forefront of new technologies. One example is their search engine, now you can simply look up a question and it will analyze it and answer it, using the top result of the search. [1]
Artificial intelligence is used non-stop in production and manufacturing of resources. There is a lot of real time problem solving that used to be up to humans. Now much of it is automated. Most things in more advanced factories are interconnected. One example is using lasers and computers to find defects on the manufacturing line and communicating them to the rest of the line. This could save future objects from potential defects by giving insight into what is causing them. Tools like this can cut unexpected downtime by 10-20%, and decrease the amount of human labor needed. The industry for smart tools is estimated to hit $250 billion by 2018. Some factories are completely automated already. These factories are called "lights-out" factories (if there is no people they don't need light). These factories still require remote monitoring. They are not failsafe, sometimes something just breaks down. [2][12]
Artificial intelligence is used often in daily life, from a simple google search to video game. A lot of news on google is actually generated by AI, especially the short financial ones. Even if the posts are not made by AI, it can help predict what you want to read about. Advertisers use prediction software like this a lot, to help predict what you might want to buy. That can also help with other financial areas, including fraud detection and market predictions. They can also protect you from spam, which often includes scams or malware. Spam often has recognizable features that AI can learn to find.
AI also helps outside of the internet, with self-driving cars or analyzing traffic patterns for example. In the United States 52 minutes a day are spent commuting to work on average. This adds up to about 26.9 billion hours of wasted time. Last year Pittsburgh tried out a new traffic system based on AI, it cut commute time down by 25%. Self driving cars could bring the time wasted down to zero. [3][4][5]
Other uses of AI can be even more beneficial. Google has already built an AI that can look at retina scans and determine if you have detect diabetic retinopathy, a cause of blindness. Other AI can show patterns in medical conditions. This can help show doctors what is actually happening, and help prevent it in the future. Some AI can help diagnose you from home without a doctor needed. [11][13]
At this point artificial intelligence is limited. It can comb through data at a rate humans could only dream of, but its decision making is lackluster. It turns out brains are extremely complex, and out of these complexities arise things like creativity, something no AI has exhibited.
Neural networks are the most promising form of artificial intelligence. They run on the same principles as our brain, but with far less complexity. They use a set of input “neurons” that connect to the next row of neurons until they reach the outputs. Each connection has a weight, which changes the input number. Neural networks learn by changing the connection weights. This gives their thinking a level of complexity not see in prior algorithm-based AI, but is still nothing compared to the human brain.
AI networks usually only run in one direction, although some do have some backwards connections. The human brain, on the other hand, is less rigid in its directioning. The connection of Neurons can go in any direction, even multiple at once. This allows the small-scale sections of the brain to create extremely complex shapes, with up to eleven dimensions. That is made possible by an average of 7,000 synapses per neuron. [6][7][8]
AI can actually complete smaller tasks much quicker than humans. Many small tasks, like recognizing spam emails, may only take a few thousand cycles of a cpu. Human reaction time can be hundreds of thousands of times slower. This changes with more complex tasks, like correctly identifying objects based on their shape. Even the best AI can be easily fooled, even by random noise.[9][10]
