2
The StoryDrive Engine
In 1995, The National Science Foundation, under the direction of the Department of Defense, sponsored a conference on the potential impact of computer games and entertainment on military strategic planning. One of their missions was to explore the use of story within computer games: specifically, the way story structure is used to present ideas within the game context.
TV and film director Alex Singer (Star Trek, Lou Grant, Cagney and Lacy) represented the Hollywood creative community at the conference. Alex brought along several Hollywood colleagues, including Richard Lindheim, then Executive Vice President in the Television Group at Paramount Pictures. The conference led to a discovery of common interests and a series of subsequent meetings between Paramount and the DoD.
Impressed by how engaging and memorable good Hollywood films could be, Dr. Anita Jones, then head of R & D at the Department of Defense, wanted to learn if movie-making techniques could be applied to the building of soldiering skills and practices. Paramount accepted the challenge. And so, as reported in the Wall Street Journal on November 11, 2001, a close collaboration began between the two entities, one that extended over many years.
Paramount and the Department of Defense began their effort with a series of research trips in order to review the kinds of simulation training carried out by the US military. They also wanted to identify an appropriate training program that could serve as a test case for the military application of Hollywood techniques. A select group of military simulation training specialists and a small team of movie, television and Internet creative people chosen by Paramount undertook the effort. Dr. Judith Dahmann, then Chief Scientist for the Defense Modeling and Simulation Office, was in charge of the DoD team. Other members included Dr. Kent Pickett of Ft. Leavenworth’s Army Research Center and Dell Lunceford from DARPA. Co-author Nick Iuppa, Vice President and Creative Director of Paramount Digital Entertainment (PDE), headed the Paramount team. Nick worked under the management and direction of Richard Lindheim and David Wertheimer, President of PDE. Alex Singer continued his involvement in the project and experienced television and interactive media writer Larry Tuch (Quincy, Carmen Sandiego) and associate producer Erin Powers were brought in to complete the team.
In order to develop an actual training prototype, a new piece of technology had to be hypothesized and defined. To help in this effort Paramount enlisted the added support of members of the Information Sciences Institute at the University of Southern California. Dr. Paul Rosenbloom was the lead USC scientist on the project.
The new piece of artificial intelligence (AI) software defined by scientists and researchers at USC and by Paramount creatives was expected to do nothing less than turn a fairly mechanical military simulation into a “Hollywood Experience.” Anyone who knows Hollywood and its workings will tell you that the not-so-secret ingredient in the best movies and television is the story. In terms of simulation training, what Paramount decided was most likely lacking in military training simulations were stories to drive them. So the software concept that was hypothesized was called the StoryDrive Engine.
Armed with a crude idea of how injecting stories into simulations would make them more compelling, memorable, and effective, the research group visited US Army and Navy bases to review the state of their military simulators. Over the past several decades the military had already committed heavily to simulation training of many kinds. At Ft. Knox, Kentucky, enormous tank simulators stand side by side, propped up on hydraulic legs that allow them to buck and gyrate as trainees simulate driving across a terrain in France. As expensive as these simulators appear to be, their cost is remarkably less than the cost of allowing thousands of soldiers to drive hundreds of tanks over vast expanses of real terrain. The savings in fuel and maintenance alone are reported to be staggering. And the transfer of knowledge is considered to be comparable.
As the Paramount/DoD survey progressed from tank, to gunnery, to aircraft, to naval simulations, one thing became very clear. At the moment of combat, when soldiers are face to face with an enemy and it is either kill or be killed, stories and all that they can bring to the experience probably don’t matter at all. If enemy soldiers are coming at you with the intent to kill, you’re not going to pay attention to their story. Your mode of operation drops immediately to survival.
So the research group began to move away from the concept of story-driven simulations for basic combat training, but at the same time they began to see further evidence of the need for stories in other critical kinds of military training, primarily in leadership, interpersonal skills training, and especially tactical decision making under stress.
In a combat situation, a soldier may have his or her hands full and should not be distracted by the bigger picture, but the leader has to consider it. Leadership decision making must be carried out with a complete understanding of what is going on. The military calls this complete understanding “situational awareness.”
And so, after a 3-month review of the many major military simulationtraining centers, the Paramount group proposed to focus on those exercises that require a broad awareness of the situation and important decision-making skills. Several such exercises were identified as possible candidates, but in the end, a large simulation called the Crisis Decision Exercise (conducted at the Industrial College of the Armed Forces [ICAF] in Washington, DC) was selected as an excellent venue for such a study.
The Industrial College of the Armed Forces shares a campus with the National Defense University at Fort McNair on the banks of the Potomac River. The student body is composed of soldiers in mid-career who show great promise and are candidates for advanced assignments in any of the branches of the military or in the Federal Government. Students also include members from US Government agencies such as the State Department, FEMA, and the USIA (the United States Information Agency).
The Crisis Decision Exercise is put on annually at the end of the school year at the college, and all graduating students are required to participate. At the time of Paramount’s involvement, Dr. Alan Whitaker, Director of Exercises and Simulations for ICAF, managed the entire course implementation.
The Crisis Decision Exercise lasts an entire week and is nicknamed the Final Flurry Exercise because it involves application of all the skills that have been taught throughout the year in one frantic, all-encompassing effort.
Late in 1997 the Final Flurry Exercise took place in university classrooms where students played the role of members of a work group reporting to the National Security Advisor (NSA) on matters of world importance. The students were confronted with a series of hypothetical international crises and were required to make recommendations to the NSA, who in turn passed on his recommendations to the president. In the exercise, the magnitude of each problem was pushed to the maximum so that students were confronted with a world in which every major international hot spot erupted at the same time.
Instructors were give latitude in the problems they chose to emphasize and their manner of presentation to the class. They had a video that highlighted the current state of the world in 1997, mentioned the trouble spots, and then opened the door to a series of classroom work sessions. Use of the tape was optional. But somehow the instructors would have to describe the state of the world, ask for recommendations, and then leave the room for an hour or so, allowing the students to brainstorm solutions. The instructors would then return and hand out note sheets, which offered updates to the evolving world crises. By the end of each day the students would have to create a set of final recommendations to be given to the NSA and then passed on to the president.
Figure 2.1 Artist conception of the workroom where students participated in the Crisis Decision Exercise at the Industrial College of the Armed Forces.
As the days went by, crises worsened and recommendations would have to be modified or changed drastically to fit the new situations that the instructor presented.
The Final Flurry Exercise was put on in 30 separate classrooms to groups of 18 students in each class. Thirty different instructors presented their versions of the exercise. To a large extent, the quality of the learning experience depended on the skill of the instructor. Instructors with a flair for the dramatic, who were able to engage their classes with their own knowledge and inventions, did best in maximizing the effect of the exercise. To put it simply, these instructors employed the elements of dramatic storytelling in their simulations. They were doing what Paramount proposed that the StoryDrive Engine do. Unfortunately, as in all teaching environments, not every instructor possessed the same dramatic skills, and so the outcomes were inconsistent.
There were several reasons for choosing the Final Flurry Exercise as a testbed for the StoryDrive concept. A major reason was that the same exercise was presented to 30 different sets of students at the same time. This provided a great evaluation opportunity. Control groups of various kinds could be set up within the population of classes to help evaluate the effectiveness of StoryDrive.
In May 1998 representatives of the Paramount/DoD team attended that year’s Final Flurry exercise. They observed its presentation in several classes and videotaped parts of the exercise. In the end, the team came away with a strategy for developing a story-driven version. Developing a full story for each scenario, populating it with realistic human characters, creating media that presented elements of the story in short snippets that could be introduced by the instructor at varying times as needed, allowing the media to be delivered to students on individual laptop computers, and building a fictional computer network with all its trappings so that participants felt that they were operating in a realistic and highly secretive environment—all these factors would enable the power of stories to increase student involvement, and engagement, and build skills that could be tested. However, in such a plan the technology itself, the AI software that monitored the participants’ progress through the story and recommended story twists and other obstacles that could maximize the participants’ engagement and challenge, would be transferred into the hands of the instructors. The military calls this kind of system “Man in the Loop” because it places a human being into the simulation and requires that that person carry out the role that could otherwise be played by technology.
The plan was in some way dangerous because it chose to focus on the underlying concept of StoryDrive rather than its technology. The danger to this approach is that the improvements to the exercise could appear to be the result of better media alone rather than an entire educational strategy with an important new technology behind it. Nevertheless, the time and financial constraints of the situation almost dictated that this approach be taken.
As a result, in 1998 Paramount adopted the most pragmatic solution, which was to test the StoryDrive concept. Working closely with instructors at ICAF, the Paramount team began to flesh out the situations and build the multifaceted characters needed to bring a Hollywood style to the exercise. The arc of each story was plotted, analyzed, and revised for maximum impact. All major characters were defined in character bibles. This is a Hollywood story development tool that requires that the childhood experiences, parents’ history, and life achievements of each character be thought through and written down in an extensive document that guides the writers in developing the characters and in understanding their motivation and actions as the story progresses.
Writer Larry Tuch presented the story’s events through various media forms, including e-mails, formal military documents, and simulated video newscasts that portrayed the characters and the results of their actions. The Final Flurry scripts and background documents evolved into a novel-length work with an equally imposing set of design documents, including story graphs that resembled flowcharts—except that at critical points there were no branches, but something more like bundles of media elements that the instructor could draw upon based on his or her assessment of the progress and needs of the participants.
Once the script and design documents were fully fleshed out, reviewed, and approved by the experts at ICAF, media production began. Drawing on Hollywood’s exceptional pool of acting talent, Paramount was able to create all the simulated news broadcasts, news anchor commentary, reports from the field, interviews with heads of state, executive speeches, and comments required by the simulation. Director Alex Singer and producer Florence Maggio cast the complex and difficult roles and created all the necessary video. News reports of world events, put together using voiceover narration by professional newscasters, were accompanied by video footage from the military’s own archives. In the end, over 600 individual pieces of media were assembled to create the difficult world situation that the participants had to address.
At the same time that the media was being created, Viacom’s interactive media group in New York created software that would store, identify, and serve up these media elements to the participants. (Viacom is the parent company of Paramount.)
The software package also provided separate interfaces for students and instructors. These were designed to look like the screens of a top secret Internet system running inside a government agency. Forms within the system allowed the participants to send messages and even formal presentations to the instructor as final products of their deliberations.
The instructor’s interface allowed the instructor to preview each media element, select it, and send it to the participants during the course of the exercise. The instructor had the flexibility to submit media elements at any time during the presentation and even to move media between days. For example, an incident slated for day 4 of the exercise could be presented in day 2, if it seemed that the an extra jolt was needed to make a point with the participants.
All media production and software development for a StoryDrive version of the Final Flurry exercise was completed by November 1998 and tested by members of the US military. Concerns were noted and a second revision of the course was created for a complete alpha test, which brought members of the Marines, Air Force, and Army to the Paramount Pictures studio lot in February 1999. Another successful test yielded more modifications, including a complete redo of most of the video elements of the simulation, to include new footage purchased from CNN to expand the news stories, and expanded performances by actors in the lead roles. At this time the whole software architecture was rebuilt by outside contractor Empire Visualization under the direction of super programmer Nat Fast.
In May 1999, the Paramount team descended on the Industrial College of the Armed Forces and, in an intense 2-week session, installed the complete StoryDrive Final Flurry Exercise in three classrooms equipped with 10 laptop computers each as well as separate instructor workstations. They tied in laser disc media projection systems to display the daily orientation video that was to be presented at the start and end of each classroom day (all other video elements were sent directly to the individual student computers as digital film clips). Paramount trained three sets of instructors to run the complete program in their classrooms and then provided a general orientation to all the other instructors. In accord with the evaluation strategy imposed by the Department of Defense, an additional 10 classes were given the complete scripts and all printed media elements as supplemental handouts to be used at the instructor’s discretion.
Figure 2.2 Sample participant screen from Final Flurry Exercise.
How did it work? On the first day in each of the three classes, the students were apprised of their roles as consultants to the National Security Advisor. They were then presented news clips that pointed out a series of international problems. The clips were identified with titles to suggest that they were news stories selected from daily broadcasts by a screening group within the CIA. The stories included reports of one antagonistic regime with a complex and difficult leader in the Middle East. Other trouble spots included a potential conflict in the South China Sea, India and Pakistan, and Latin America, where severe droughts were fomenting a humanitarian crisis.
The class was asked to review a series of background documents that had been provided on their laptops. These included intelligence reports, dossiers, political maps, historic studies, and other deep research. The class then had to discuss the crisis points and to formulate a series of recommendations for the president. The president’s own political fortunes were factored into the equation as well. The news stories included information on the upcoming election year. The president’s most likely opponent was the governor of a large southwestern state, who was highly critical of the president’s inability to deal with international issues. All these stories were laid out on the morning of the first day, and then, gradually, as new information was disclosed in messages and breaking news broadcasts, the situation grew worse and worse and worse.
Two years before 9/11, the threat of international terrorism was a major element in one line of the story. Hypothetical terrorist attacks were detailed in several news reports. Over the course of the week, charming but nefarious adversaries, opportunistic despots, drug lords, and the weather itself conspired to challenge the participants in ways that were made crystal clear to them through daily media reports and urgent messages. If no one noticed the messages the course of the conversation might go on unchallenged, but new events were almost always identified at once and often greeted by surprise and excitement that brought a charge of electricity to the deliberations.
At the end of the business day the NSA (who existed only on video) would come back to critique the class recommendation, and then the fictional president would address the nation concerning the major international issues that the class had been discussing. If the members of the class were on their toes, his speech would echo their recommendations.
The president’s speech was actually constructed by the instructor from a large database of prerecorded video clips, each clip offering a different response to a specific crisis. The original content of the clips was based on input from the ICAF advisors who came up with all the possible recommendations they could imagine being given by the class. The database of video statements was stored on laser discs and accessed by the control software that was part of the instructor interface. If the class came up with recommendations that no one had thought of before, the instructor had the ability to create a special text memo from the NSA that would provide very specific feedback explaining why the president did not include that particular recommendation in his speech. But more often than not the predictions of the instructors proved to be accurate and the crisis responses that the students recommended matched, sometimes almost to the word, the prerecorded statements by the fictional president. Students in the class were dumbfounded when they heard the president repeating the words of their recommendations in his prerecorded speech.
After the first day, news of the StoryDrive version of the Final Flurry Exercise spread to the other classrooms. Most students were very interested in the new approach. Many wondered why they were not able to see the same presentation of the materials. By the end of the full week the interest in the story-driven simulation was so great that ICAF determined to spread that approach to the exercise to the full student body the next time the exercise was presented.
The attitudinal surveys that followed confirmed the fact that the story-driven approach did a better job at engaging the students and allowing them to focus on the difficult issues involved in the simulation.
The project was regarded as a great success at ICAF and in the Department of Defense, and it was generally conceded to have proven that story-driven simulations can be effective in military training. What had not been demonstrated was a working system that used AI to maintain the story and the high level of drama needed to test the ability of the participants to make decisions under stress.
Fortunately, other departments within the US military became interested in carrying that effort forward. All of which lead to the second of the three major projects that Paramount Pictures carried out in pursuit of story-driven simulations.
SUMMARY
The Department of Defense became interested in creating a simulation-training program that used storytelling techniques to make military training simulations more compelling and effective. Paramount Pictures accepted the challenge and with the DoD embarked on a survey to find appropriate subject matter for a study of the concept. Leadership and crisis management simulations were identified as the best candidates and the Final Flurry Exercise at the Industrial College of the Armed Forces became the focus of the effort. Paramount developed a complete media package that brought storytelling techniques to the exercise. The exercise was presented to three of the 30 classes conducting the Final Flurry Exercise and was judged to be so successful that the leaders of the school adapted the approach for use across the board.