Thursday, August 8, 2013

The Debate Over Artificial Intelligence

I've been reading quite a bit about artificial intelligence (AI) lately, and I'm fascinated by the polarisation surrounding the topic. Today's breed of smart computers and robots can beat humans in chess and outperform financial traders in investing stocks. They can also auto-drive cars, detect tumours, and, in the case of drones, drop bombs without any human intervention required. Like all technology, the potential uses are endless and AI can be used for "good" or "bad" purposes.

I see the advantages that artificial intelligence can offer. Smart computers can carry out rote work that we all loathe, and process mass amounts of information, enabling humans to synthesise, analyse, and creatively design solutions. Robots can also provide an opportunity to support humans in becoming more efficient and effective. This is critical in the fields of development, rural medicine, and disaster relief, where coordination is difficult and resources are stretched.

But there are several key debates that we must iron out as we move to more AI. (I will not discuss the concern that, as computing power increases, evil robots will take over the world. This is a great leitmotif for science fiction movies but not exactly relevant for today's AI.) The two biggest concerns around AI today are morality and humanity. Robot ethics is an emerging field. If cars can drive themselves and machines can prescribe medicine, how do we ensure that these robots behave in a moral way? This isn't as straightforward as it seems. Imagine a drunk cyclist swerving in front of a self-driving car. To the left of the car is a sidewalk with a child walking; to the right is another car. What does the self-driving car do in that situation?

Another fear of AI is that it will make us less human. Many technologists describe a near future where humans implant chips into their bodies- enabling them to check their emails and search Wikipedia without the need for a device. Google Glass is seen as a gateway to this new way of accessing information. The debate here centres around our sense of humanity. What does it mean to be human if we suddenly integrate machines into our bodies? Will this evolution turn us into something else? Or is this just a seamless extension of our smartphones?

I love Shyam Sankar's view of cooperation between man and machine, as opposed to man vs. machine. With the aid of smart computers, human can make better decisions and achieve exceptional outcomes. My favourite example from his TED talk is a human + machine chess tournament. Humans entered the tournament with the assistance of machines. The winner of the tournaments was not a chess master with a super computer, but rather a team of two amateurs working with three regular laptops. When man and machine work together effectively, exceptional outcomes are possible.

I hope that a man with machine paradigm is the future of AI. But as we programme robots to take humans out of the equation, we need to find ways to ensure moral outcomes (both by programming moral robots and by understanding the consequences of virtual warfare using drones.) And we need to carefully consider the possibilities of infusing robot-like technologies into our bodies. I am optimistic about the future of man and machine, and I believe that healthy debate will be key to guaranteeing positive outcomes with AI.

Image courtesy of Interestingtopics.net

No comments: