/Artificial Intelligence: Nature of the Possible (via Qpute.com)

Artificial Intelligence: Nature of the Possible (via Qpute.com)

The readers of my series of posts on Artificial Intelligence will probably notice that the potential impact of AI technology is sometimes presented with great enthusiasm and also with considerable caution. Some of this schizophrenia is due to the different outlooks of the capabilities and the technology that tend to regard the world from opposite ends of the enthusiasm-pessimism spectrum. Beyond personal viewpoints, however, is the uncertainty associated with any new technology. We all can agree that Al technology will undoubtedly revolutionize the productivity of many components of North American and global industry as AI comes into more general use. On the other hand, it is not necessarily true that the introduction of this technology will be either easy or cheap. After all, software is software, whether Al or not, and all the problems associated with the development and maintenance of complex software still remain.

Rapid prototyping technology and knowledge engineering methods can be used to improve the specification process for large and complex systems, whether Al technology-based or not. Also, many well-bounded areas of human decision making should benefit from the use of rule-based systems. If the decision space for a system is not well-bounded, it is more likely that an intelligent assistant system will be appropriate. Going beyond these limits to the areas of human capability in decision-making under uncertainty or learning from experience seems unlikely in the near future. Such capability in a computer will probably not be feasible except in limited demonstrations until both the hardware and software for massively parallel computing or even quantum computing are commercially mature.

With that said, we are seeing some amazing leaps in the capabilities of AI and Machine Learning.  However much of this signal is also being lost in the noise of a lot of technologies and applications that claim to be using AI when they are just using some basic algorithmic solutions that don’t really do any ‘thinking’ or act in any way ‘intelligent’.  Just because you can create a complex web of ‘if this then that’ processes does not mean you have artificial intelligence – all you have is a crazy sit of Plink style decision trees that will likely just frustrate your users and not deliver any real value to the industry.

This is a syndicated post. Read the original post at Source link .