Let’s talk a little bit about Artificial Intelligence today. My take on it is that, given continued technological improvements and research, one day we will engineer an artificial intelligence more “intelligent” than humans. I don’t know when this will happen, or what form such an intelligence would take, but it is inevitable. This moment is commonly described as the singularity point. It is called that because we have no idea what will occur from that moment on, the idea being that once machines surpass humans in intelligence they will be able repeatedly improve on their own designs in new unforeseen ways, leading to an exponential growth in intelligence.
Where we go from there is a mystery. It could be disastrous for humankind, or really beneficial. Anyways, we’re still quite far off from that faithful day. There are people who focus solely on the type of AI research that would lead to the singularity (and you can read all about it here), but the field of AI is vast, and research is being spear-headed in many other areas. It’s the combined results in those fields which will continue to improve our lives. Let’s take a brief random tour:
- In the virtual world Second Life, researchers from Rensselaer Polytechnic Institute have engineered an Artificial Intelligent agent which has the mental reasoning ability of a 4 year old child, and continuously learns from his environment and interactions.
- Arstechnica have done a series on the future of self-driving cars (partially inspired by the thorough research done by EFF chairman Brad Templeton). In fact, the push for automated vehicles is becoming more pronounced every day: researchers at Berkeley have demonstrated a highly accurate automated guidance system for buses, and the EU has reserved a radio frequency band solely for the use of communication between cars and the road infrastructure. This topic could use entire post on its own.
- Meanwhile, a PDA-like device is being developed by DARPA (you know, the guys who brought you the internet) for use by American soldiers in Iraq and Afghanistan, which will automatically translate English phrases into the local dialect in real-time. In fact, this article bills the technology as being straight out “star Trek”. I’m telling you, the future is here. Science fiction becoming reality. It’s fantastic.
I’ve gone the long way around to get to the main topic of this post. I wanted to point you to a series of articles this week on a nearly successful Turing test. A Turing test is used to determine if a computer is able to demonstrate human-like intelligence during five minutes of questioning from a number of judges. In order to pass the Turing test, and AI program needs to fool 30% of the judges. The annual Loebner Prize contest held this week seeks to reward AI entries which fare the best in these Turing tests. And this year, one program came tantalizingly close.
Passing the Turing test isn’t the threshold for true artificial intelligence, but just another step along the way. What is clear is that every day, our machines are being made more intelligence.
Related posts:
Leave a Reply