The early days of the new century were tough on the IT industry. The dot-com crash hit just about every market sector, and the field is still recovering. Yet the artifical-intelligence industry largely dodged the crash, because it didn't primarily focus on the Web and Web-related areas.
Instead of exploiting the current Web, many AI companies are changing it. That's why they didn't need to shift gears in their ongoing research when the bubble burst. Specifically, they're developing the Semantic Web. This attempt to create a new generation of Web applicatons brings semantic technologies into a form compatible with Web architectures.
Experts say this could create a multibillion-dollar market driven by a large degree by enterprise aplication integration. It's the result of a convergence between Web developers and AI researchers. The Defense Advanced Research Projects Agency and the European Union's IST program both have invested in it.
Researchers are developing Web-based standards for the ontologies and business rules used in e-commerce, supply-chain and transaction management, and homeland security and counter-terror applications. Oracle, Nokia, Fujitsu, and IBM are developing or using Semantic Web products. So are millitary contractors, including Northrup Grumman, McDonald Bradley, and Lockheed Martin. Companies in the biotechnology sector have expressed an interest as well.
Meanwhile, AI is flourishing in robotics. While traditional robotics have focused on industrial applications, AI has shifted to the concept of the ?autonomous? robot. Rodney Brooks of the Massachusetts Institute of technology is leading the charge. He helped form iRobot, the creator of the popular Roomba vacuum.
The Roomba was the first non-toy robot to successfully break into the commercial sector. Over 1.5 million Roombas have been sold since its introduction in 2002. Meanwhile, iRobot has developed the Packbot robot for the military. It has proved its worth in Afghanistan, and it's now being used worldwide.
The success of the Roomba and the Packbot, in addition to Sony's AIBO robotic dog and countless imitators, demonstrate the potential for a robust home market. The extension of this ?small robot? technology to larger devices, particularly in transportation, should become more exciting over the next decade. Just look at the success of the autonomous vehicles in DARPA's annual Grand Challenge competition, which navigated a 142-mile course through the desert.
Another interesting area is humanoid robotics. Driven mostly by Japanese firms, humanoid robots mostly have been used to show off the ?futuristic? design and manufacturing prowess of commercial giants like Sony and Honda. But these companies make a compelling argument for the future of these products.
Much of our environment and the machines we use were ergonomically designed for the human body. Robots that could drive a standard car, load a regular dishhwasher, pack and unpack items in a department store, or perform any kind of dangerous or onerous manual labor must use a human-like design. Re-engineering these environments, manufacturing lines, and warehouses would be impractical.
In addition, expect the need for these ?support? robots to grow as the populations of the world's most developed countries ages. These robots will need to interact with the standard home instead of their owners, who may not be able to handle routine domestic chores anymore. The success of the Japanese effors has spawned a large number of companies and academic spinoffs looking to enter and ultimately conquer this potentially huge market.
There are some other exciting areas to watch, too. Microsoft Research Laboratories continues to explore the role of probabilistic AI techniques for e-mail screening and corporate knowledge management. Google, Yahoo, and other companies are looking at machine learning technology coupled with language processing as important technologies in the ?intraweb search? area. And, the military is exploring the integration of AI techniques with service-oriented architectures as an approach to both legacy system integration and self-adapting ?cognitive systems.?The current directions in the Semantic Web, robotics, and these other areas are significant departures from the AI of the 1980s. As an advertising slogan might say, this is not your father's artificial intelligence. If you haven't looked at AI recently, it may be a good time to renew your acquaintance.