Slashdot Mirror
Software Agents Can Help Time-Stressed Teams
Roland Piquepaille writes "Penn State researchers have developed software agents which can help human teams to react more accurately and quickly in time-stressed situations than human teams acting alone. According to this news release, the software was tested in a military command-and-control simulation. "When time pressures were normal, the human teams functioned well, sharing information and making correct decisions about the potential threat." But when the pressure increased, the human teams made errors who would have cost lives in real situations. The decisions taken by agent-supported human teams were much better. Now, it remains to be seen if this software can be used in other stressful situations, such as for emergency management operations. Read more for other details, references and illustrations about this project."
Is this the beginning of Skynet?
What would be interesting is if the software agent could be programmed/adapted for each individual team. The team uses those single word/syllable communication methods, and the computer is made to understand them (during training exercises). Then, in high-stress situations, when that word/syllable is uttered by one of the teammates, the communication is put into the software system. The computer knows what all the team members collectively knows. Then it is able to make recommendations on course of action that the individual team members might not be able to make, due to the way humans are affected by stress.
The way I see this, it doesn't replace good people, good training, and good exercises, it merely assists the people when they need it. I think it could be interesing. It would only work, however, if it were completely transparent. No little application running outside of their normal software suite that pops up and asks for confirmation. Instead, it should be built into the software they normally use and should alert the user only when it has a recommendation. All users are notified simultaneously, or it could be given only to the team leader, who could have final say over whether the computer recommendation is valid or not.
I can even see this kind of system reducing the stress for teams in these situations.
Lack of eloquence does not denote lack of intelligence, though they often coincide.
It all boils down to how good the information you are getting, and how good your training is.
There's not enough information in TFA to determine how much work the agents are actually doing... whether they are making decisions on their own, or simply hiliting and collating what appears to be relevant information.
If one of the humans has to collect information from several sources and collate it, what does it hurt for the computer to do that, and include a recommendation... enough training and the recommendations just become another datapoint which you intuitively incorprate into your actions.
However, if information is hidden or removed from the human, or the agent is making decisions at too high a level, then I'm very worried.
I have re-built several helpdesk and incident response systems, and the the main productivity enhancer I bring to the table is having the server perform some basic analysis of whatever is about to be displayed.
Simplistics attitudes like this get in the way of real decision making.
What if by allocating resources to this project, the project to build a resource allocation system for medical personnel is scrapped?
Anyone who says "If it saves only one life..." has turned off their brain. How about this... We take your house and turn it into a homeless shelter (you included). It will save several lives. If we do it to 100 people in 50 cities in colder climates, we can save hundreds of lives in one winter, for almost no money.
For example, just out of highschool I took a summer job as a hooker (insert jokes here). I was a hooker on a log salvage operation in the mountains on the west coast, (Canada). As a hooker my job was to catch a hook (thus hooker) on the end of a long steel cable and attach it to thickly braided nylon cord, braided at the end of the cord into an eye, wrapped around a section of log to be salvaged. (The fun part of the job was ridding the hook from one salvage site to another... I would wrap my legs around the hook on the end of the cable and the chopper pilot would fly me from one mountain side to another. Very illegal, but oh what a rush.) At one site the chopper came in and titled away to "throw" the hook at me. The idea was that the chopper tilted away from me, guiding the cable hook to me; I would catch the hook in one hand, with the eye of the nylon cord in my other hand, snap the hook into the eye of the cord and the chopper would take off, still tilted away from the mountain side, taking the log section away to a dump. It called for speed, concentration and preparation. The nylon cord had to be looped just so on the log section, so that it wouldn't tangle. On one occassion the chopper came in, tilted, and feed me the hook. I snapped the hook to the cord and threw my hands back signaling the chopper to fly off. What I hadn't seen was that someone had made the cord too long and had left it looped on the ground. My right foot was in the loop. The chopper took off at about 50 Km, the loop began to close. Between the chopper going away at 50 Km and the log section weighing in at 450 Kg my chances of survival were negligible. Even if I had just lost my leg the nearest hospital was an hour plus away by chopper.
Time went away (I've no other way to describe it). There was just the loop snaking up my leg. My mind was crystallized, there was no thought, no mundane awareness. Awareness of my body was gone.
I did a perfect back flip, pulling my body up and away from the closing loop, landing on my shoulders, then tumbling back upright. The chopper took off cleaning jerking the log section away. I'd taken a few tumbling classes in jr high, but wasn't anywhere near the training necessary to what I'd done. The sense of purity and oneness such situations bring is highly addictive.
Under pressure some crack, others look to the alpha members of the group to OK their actions. Some exit time and do what needs to be done. For the latter group no software is going to keep up.
In the alternative, I watched a documentary on Vietnam, wherein a US officer said the lesson that stayed with him was never to send a man, where a bomb or a bullet can go. I think the effort is to follow the 5Ps (proper preparation prevents poor performance) and, as much as possible, plan for events such that protocol substitutes for heroics, and, in this guise software can reinforce protocol.
"Academicians are more likely to share each other's toothbrush than each other's nomenclature."
Cohen