Can complex parallel universes created using Artificial Intelligence (AI) really prepare U.S. soldiers for conflict in countries like Afghanistan?
He was getting nowhere. The more Private Neil Lambert of the United States Army quizzed the headman on why there were outsiders in the village at night, the more opaque the Pashtun chieftain's answers became.
While the stalemate was maddeningly frustrating, Private Lambert was beginning to understand his interlocutor's truculence and he had a theory. He knew from his pre-mission briefings that the Afghans in this village had a tendency to gossip and rumours spread quickly.
How could you blame them? After a decade of Western occupation, trust had become a scarce commodity in this barren land and word-of-mouth was as good as any other basis on which to form opinions. And Private Lambert had messed it up — he'd had a falling out with his interpreter a few days ago and that man had, it was becoming increasingly evident, gone on a rumour-spreading rampage.
Until today that had only meant that the Private's investigations into the unusual nightly visitors in the village had drawn a blank. But now the situation was deteriorating and the headman was getting aggressive. Even worse, as the Private tried his best to keep calm and deflect the headman's angry allegations, a mob was beginning to form around them.
Then it happened.
Out of nowhere a Kalashnikov rifle materialised in the hands of one of the villagers. Private Lambert had two choices: first, he could quickly swing his own rifle into position and face down the armed villager; or second, he could back down, strike a more conciliatory note and keep the focus on the headman. In seconds he made the choice and was about to respond when, suddenly, everyone froze.
The headman was frozen in mid-sentence, about to utter an angry profanity at Private Lambert. The gun-toting Pashtun who was advancing threateningly on the soldier stood as if paralysed, unable to swing his weapon upwards. The rest of the agitated mob was in suspended animation, as if a collective “pause” button had been pressed. The only man who could move was the American soldier, and he gasped in surprise as a voice boomed from the sky: “That's the end of this session Private — you have two hours to plan your responses and come back to your station.” Then everything went black.
While Private Lambert's surreal experience might be mistaken for a remake of the movie “The Matrix”, it is not. It is real and U.S. Department of Defence personnel engage in such virtual-universe exercises on a regular basis, thanks to the work of a top arts and technology college at the University of Texas at Dallas (UTD).
At this powerhouse of animation technologies, the aim is to go far beyond the industry's most common application — entertainment, usually in the form of Massively Multiplayer Online Roleplaying Games Programmes, or MMORGP. “We don't believe games should be purely for entertainment, but we do believe they should challenge the living daylights out of you,” says Dr. Thomas Linehan, Director of Arts and Technology (ATEC) at UTD.
The UTD team of top animation specialists and even some “young Turks” with a passion for programming — in the words of Dr. Linehan — engages with government agencies such as the Pentagon and the National Institutes of Health to create virtual-reality worlds of vast depth and specificity.
It is these complex parallel universes based on Artificial Intelligence (AI) that soldiers deploying to Afghanistan and nurses specialising in paediatric and neo-natal care rely on to get the training that a real world classroom could never give them.
For example, the award-winning programme aimed at acclimatising new recruits to the army to the alien surroundings of a remote Afghan village is called the First Person Cultural Trainer (FPCT). In this game, the soldier enters the Afghan — or Iraqi, as the case may be — community from the first-person point of view.
While the soldier does not initially know about the community, how the people feel about him, who the powerful characters in the village are, or even how to interpret their body language, “the goal is to move through the community, try to understand social structures and issues, then address those issues and work with the community to affect missions,” says Dr. Linehan.
The FPCT game is supported and sponsored by the U.S Army Training and Doctrine Command whose head, General Martin Dempsy, said in a personal letter to UTD, “Your magnificent work in developing this culturally-based, cutting-edge capability will help to address critical training priorities within stability operations.”
It is not an exaggeration to describe the FPCT as “cutting-edge”. It requires not only advanced code-writing skills but also a creative impulse that does not shy away from unusual approaches, for example, the use of actual Pashtun individuals recently arrived in the U.S., who regularly provide the UTD team with live demonstrations of their body language, their intonations and accents, even their views on the Western occupation and their attitudes towards U.S. soldiers.
The UTD team then “maps” every such physical and psychological attribute of the Pashtun man or woman — both genders are analysed and corresponding virtual characters created — sometimes using motion-capturing technologies used in Hollywood films.
The AI aspect of the programme, which makes it unique, arises when feedback loops based on the soldier's interactions with the villagers are required, so that the villagers begin to respond and react to the soldier over time.
There is even a gossip programme, which might have produced some of the nerve-wracking results that Private Lambert faced. “If the player does not interact properly with them, the villagers discuss his behaviour among themselves. Some individuals in the village have more clout than others,” Dr. Linehan explained.
Substitute for understanding?
While this alternate-reality scenario would likely hold inherent appeal to anyone interested in gaming, social scientists might wonder about several fundamental questions that it raises. Can the true depth of Afghan village reality ever be captured by a computer programme, even one that relies on the inputs of real-life Afghan villagers? Does the West's bitter failure to stabilise Afghanistan after a decade of occupation suggest that such a formulaic approach to cultural sensitisation of soldiers has failed?
Army statistics cast an illuminating light on these issues. Based on evaluations of the efficacy of the FPCT programme, they suggested that a soldier who had undergone sensitisation was 50 per cent less likely to generate “kinetic energy” around him — a euphemism for shooting someone or being shot – than one who did not have such training.
Regarding the importance of cultural sensitisation per se, a majority of the U.S. soldiers who died in Afghanistan did so during the first six weeks of their tour. “So what is it that an American soldier knows after six weeks and one day that he did not know until the previous day?” Dr. Linehan asked.
Innumerable training benefits also emerge from other virtualisation programmes built by the UTD team, which includes Dr. Marjorie Zielke, Assistant Professor at the ATEC programme, and Gopal Gupta, Head of the Computer Sciences Department. Together, the ATEC team has led the development of a second project that guides nurses through realistic simulations of paediatric care.
In this programme, called NursingAp.com, nursing students find themselves in a virtual neonatal care environment in which they face a variety of challenges from how to operate the “virtual ventilator,” to dealing with the stress of answering questions from the patient's family or “feeling” fatigue of from the intensity of the job.
However especially in the context of the FPCT, the deeper question relates to the potential of such virtual-reality-based training programmes to move beyond the battlefield and be applied to government officials in a position to prevent the outbreak of conflict in the first place.
“While we are happy with the work we are doing with the Department of Defence, our ideal approach would be to someday work with diplomats in the State Department,” said Dr. Linehan. If his dream is realised it could herald a new era where the lines between virtual and real worlds are blurred in the interests of conflict resolution.