AI, or artifical intellience, is one of the most mis-used terms in technology today. Let's break it down
- AI: a computer that can learn. There are no computers that can do this independently today.
- Machine learning: a cognitive service very focused on a skill (vision, speech, language, search, etc). These are very good at doing one thing - extracting text from an image, converting text to speech, or understanding the intent of a sentence. These are the building blocks of AI, and tyed together are very powerful - for example searching images and then using image recognition to verify the images are what we expect.
- Deep learning: can learn a very focused skill. For example, picking out text from an image
What is the state of AI today? There is no artifical intelligence. Most AI is really just machine learning and statistics. You take a picture and machine learning uses statistics to tell you what is in it, what the text reads, and facial expressions.
Note that AI is only is good as the data passed into it to learn. The data needs to be carefully filtered and organized. Garbage in == Garbage out. Where AI really excels today is with cognitive services. Image recognition, text<->speech, image->text, etc. These are very focused, well honed algorithms that work well.