How Artificial Intelligence May Change How Agencies Recognize a Terrorist Threat
The idea of artificial intelligence has been a science fiction concept in books and movies for many generations. Humans have always had the dream of developing computers and robots that could think, behave and respond to the world in the same way that humans do.
As computer technology continues to advance and eventually shrink into the nano-scale while at the same time increasing in computational power, the idea of reproducing the complexities and abstract thought processes of a human mind slowly become more practical.
My most accounts, the field of Artificial Intelligence research started at Dartmouth College in 1956, when computer experts of the time – men like John McCarthy and Herbert Simon and their students – produced computer programs that could perform mathematical and logical tasks that most people thought only humans were capable of. (1)
By the 1960′s, computers and the idea of artificial intelligence soon entered into the realm of national security, since – as with most fields of technological research – both the United States and the Soviet union became concerned that the other side would establish an advantage over the other.
Artificial Intelligence as a National Security Concern
AI is no less a national security concern today than it was back then.
However, today, the “enemy” is the cyber-terrorist, or the foreign state with an intent to cause damage to Western computer networks, or otherwise influence the integrity of the growing, interconnected system of data communications and data storage.
This concern is documented in a 1962 SECRET CIA document describing several secret Soviet documents leaked by a Soviet source. The documents were all marked with code name IRONBARK and came under Richard Helm’s tenure as CIA Director.
The documents detailed a number of Soviet SECRET articles from the June 1961 issue of Collection of Articles of the Journal Foreign Military Affairs published by the Soviet Ministry of Defense.
Soviet Use of Artificial Intelligence
The one that included mention of artificial intelligence was Colonel P. Savinskiy’s article titled “Methods of Achieving Tactical Surprise in Ground Troop Operations”.
In that article, Savinskiy wrote:
“For this all possible means of deceiving the enemy will be used: setting up of mockups; simulated activities of communications means and radiotechnical stations, which service subunits using nuclear means; creation of artificial intelligence indicators of the location of nuclear means, such as, for example, evacuation of the civilian population and reinforced security in specific areas…”
Within this context, the term refers to a careful analysis for areas potentially requiring civilian evacuation and enhanced security. However, what the use of the term in this document shows is that the concept of “artificial intelligence” was in use within intelligence circles and academia long before it became a commonly used term throughout society and in the media.
Intelligence Community Interest in AI
Artificial intelligence was always a concept under study and development by U.S. academia and the intelligence community.
In an April 3, 2001 speech prepared by John C. Gannon for the National Security Telecommunications and Information Systems Security Committee, Gannon stated:
“These changes [pace of change in information technology] could improve processing power, information storage, and bandwidth enough to make possible application of advanced software technologies-such as artificial intelligence-to cyber warfare.
-> Such technologies could provide the defender with improved capabilities for detecting and attributing subtle malicious activity, or could enable computer networks to respond to attacks automatically.
-> They could provide the attacker with planning aids to develop an optimal strategy against a potential target and to more accurately predict effects.” (3)
These comments revealed an intense interest in using cutting edge Information Technologies to detect and analyze “subtle malicious activities” within the volumes of data collected by the various intelligence agencies.
Terrorism Sparks Increased Interest in AI
Because of the existing interest in AI, the terrorist attacks of 9/11 in New York city only served to fuel that interest and intensify efforts to make better use of IT technologies and to focus on further developing the cutting edge of artificial intelligence – particularly as applied to intelligence collection and analysis.
By 2003, the National Science Foundation had established the Artificial Intelligence and Cognitive Science (AICS) program – focused on “advancing the state of the art in Artificial Intelligence and Cognitive Science.” (4)
This increased focus on IT and artificial intelligence was not limited to the intelligence community.
By March of 2003, law enforcement agencies around the country begame linking up to a nationwide system developed in part by the National Science Foundation’s “Digital Government” program.
The program, called COPLINK, used:
“…techniques from artificial intelligence and other fields to help sniff out the faint trails woven through vast databases, potentially across police divisions or among agencies, and provide investigators with leads for their cases.” (5)
The AI aspect of the system was rather remarkable. While previously a human analyst would need to sift through volumes of data from the various agency databases and other systems, the new computerized system coulc perform its own analysis – detecting specific patterns within the volumes of data.
Since 2003, developers have continued developing and improving the “detection algorithms”.
While this is good news for law enforcement and bad news for criminals, some civil liberties advocates question just which “agency databases” the system can access, and whether some of the data analysis might trample upon the privacy and civil liberties of regular Americans who pose no threat to national security.
Develpment of AI for National Intelligence Continues
As the field of computer technologies continues to advance into the nanoscale and computing power grows at an exponential pace, the interest of the federal goverment in making use of artificial intelligence hasn’t waned.
Evidence of this comes from a March 29, 2012 Fact Sheet from the Executive Office of the President which details the importance of “Big Data Across the Federal Government”.
One of the top programs promoted by that fact sheet specifically mentions artificial intelligence.
“The Machine Reading program seeks to realize artificial intelligence applications by developing learning systems that process natural text and insert the resulting semantic representation into a knowledge base rather than relyin expensive and time-consum?ing current processes for knowledge representation require expert and associated knowledge engineers to hand craft information.”
Clearly, this indicates that the Cold-War era of human intelligence analysts sifting through documents for select pieces of critical intelligence information is fast becoming an artifact of the past.
In the future, supercomputers buried deep within the Pentagon will become even better at processing mind-numbing volumes of text, video and audio data in a never-ending effort to spot specific trends and potential threats to national security.
The White House Fact Sheet indicates that the U.S. Government has not yet brought that scenario to full reality, but it is only a moment of time before the science-fiction of true artificial intelligence is fully realized.