Artificial intelligence is becoming part of everyday life. As you type a search query in Google, an algorithm that learns what searchers are looking for helps you find your destination quicker. But the biggest shift AI may usher into the world is to change the face of warfare.
"In the past, security work was really more arms and weapons," said Hsinchun Chen, director of the Artificial Intelligence Laboratory and McClelland professor of management information systems at the University of Arizona. But, Mr. Chen said, over the past decade, the face of national security has changed, and now depends more and more on information.
Indeed, experts predict developments in artificial intelligence will only accelerate, and allow countries to sift through e-mails, cellphone conversations and photographs to find terrorists, for instance, and guide unmanned drones. They will change the nature of war.
Here's a glimpse into that future.
The most attractive feature of artificial intelligence for governments is that it exceeds the limits of humans. AI applications can monitor communications around the clock to hunt suspects or detect terrorist plots.
Applications may one day allow wanted individuals to be identified any place, any time, by analyzing security camera footage. Cameras may be able to detect unusual patterns of behaviour, which could then alert security staff. Such systems are already being used in prisons, for instance, to alert staff before fights or riots break out.
Facial recognition is also generating significant attention. While such software already exists, it is relatively primitive - it can perform poorly in low lighting or if an individual changes his or her hair colour, said Adele Howe, computer science professor at Colorado State University and executive council member of the Association for Advancement of Artificial Intelligence.
In the future, software could identify individuals regardless of the lighting or whether their appearance has changed. Such developments could expedite time spent at border crossings and help authorities accurately identify suspects involved in crimes.
The use of AI for security purposes is often seen as a tension between the excitement of new possibilities and fears over the loss of privacy and personal freedoms.
One of the most pervasive and controversial applications of AI for security purposes is the monitoring of communications. Computer programs can identify patterns or particular words in e-mail or cellphone conversations that may signal criminal activity.
The massive amount of information that is exchanged online and in phone calls every day makes it impossible for humans to spot potential risks. But computer scientists have devised ways to identify patterns or keywords in seemingly infinite amounts of data. This technology is what helped U.S. intelligence officials locate Osama bin Laden.
Prof. Chen and his colleagues at the Artificial Intelligence Lab have received U.S. government funding to track individuals on the so-called "dark Web" - websites and networks that are impossible to find through conventional methods, and which have become home to criminal groups and terrorists. He believes this holds major promise for tracking terrorists and understanding how individuals become radicalized. "You look at terrorists, look at their content, ideology, violence, information, warfare," he said.
But experts also warn these developments may come at a serious price - the elimination of basic privacy and the pooling of ever greater powers in the hands of the military and governments.
These implications must be addressed urgently, Prof. Howe said. "It's the way information can be used," she said. "Information can be used for good reasons and information can be used for not so good reasons."
For instance, U.S. intelligence may have computers look through e-mail or listen to phone calls to find criminals, but average citizens would lose their right to keep their communications private. This also raises the possibility, however remote, that private, sensitive information could somehow become public or be exploited without a person's consent.
On the ground
In many respects, no matter how sophisticated AI applications become, they won't replace on-the-ground warfare. But new developments could mean fewer human soldiers are put in danger.
One of the most visible ways AI is being employed in a security context is the growing use of drones or unmanned vehicles to limit military casualties. The U.S. has ramped up their use in the Middle East significantly under President Barack Obama. The country's growing use of drones has generated hostility, especially in Pakistan, where critics say numerous civilians and very few criminals have been killed in attacks by drones.
But despite the backlash, it doesn't appear drones will disappear: Earlier this year, a Florida-based company working with support of the Pentagon unveiled a new aerial drone that is the size of a hummingbird.
There's been an "incredible increase" in the number of unmanned vehicles used by the military, Prof. Howe said. "It's made a major difference in how we're able to go in, how many people have to be put into harm's way," she said.