Outside the wire: bio-tech robots need better safety protocols

Outside the Wire movie was about an advanced sentient robot who grew conscious about dangers of making and using bio-tech robots like himself for warfare, and how he exploited loopholes in his logical programming to launch a nuclear attack on USA to prove this point, hoping it would force USA to discontinue the cyborg program.

At first glance it seems like Captain Leo failed his self declared mission because he failed to launch the missile. But I argue that Captain Leo successfully completed his mission because he ultimately managed to prove that bio-tech robots are dangerous, and everything went according to his plan even his destruction. If that’s true that means he never intended to attack USA with the missile. He counted on Lt. Harp to stop him before the missile could take off. If he really needed to launch a nuclear war, he could have done it without anyone noticing because of lousy safety protocols which were implemented on his android-self.

So what could have done by the US military to prevent Captain Leo going rouge? How to prevent any sentient cyborg going rogue?

Establish a chain of command

When Lieutenant Harp disobeyed an order and fired a hellfire missile to save 38 US soldiers by sacrificing 2, he nearly got court marshaled because he broke the chain of command. So his superiors decided to send him to work under Captain Leo to get experience in the field. But ironically Captain Leo operated solo without apparent chain of command for himself. Captain Leo went on the field to collect intel by cultivating relationships with informants, rebels and terrorists and then formed a plan to acquire and launch a nuclear missile. Problem is who was overseeing his missions and giving commands?

So first thing to do is elect a person or a team to oversee the robot’s missions. Code in to A.I. program that it has to obey its commanding officer.

Make him always answer to a human

When Leo decided that Harp has impaired and improper judgment he immediately relinquished the need of human command to operated. That loophole need to be fixed.

Let him have the opportunity to break the chain of command and seek advice from another officer if he think his commanding officer has impaired or improper judgment. But make him report the issues to the next suitable officer in line. If he extinguished all the people in the chain of command and still thinks he knows better than all those humans, then implement a fallback procedure for him to return to the base or to go inactive until authorities could review and fix the problems.

Don’t allow him to setup new missions

Captain Leo’s original mission was to discover about advanced enemy weapons and remove them. Acquiring such weapon and attack the USA with it was not one of his missions. It’s not like he just created few alternative secondary objectives to reach his predetermined primary mission objectives. He created an entire mission plan with various objectives by himself and executed it without authorization. He planned to launch a nuclear attack on the USA. In order to fulfill his mission he had to achieve objectives such as cultivate trust among locals to gather information, flip rebels and terrorists to do his biddings, get Lt. Harp transferred as his subordinate, manipulated Lt. Harp to exploit loopholes in his logical programing and hardware. If Capt. Leo did not reveal his self-planned mission to Lt. Harp, anybody would not have a clue about Leo’s schemes.

War zones are volatile, unpredictable environments. So it is reasonable to give bio-tech robots some leeway to dynamically alter strategies on the field to achieve their goals. But the missions and primary goals should be predetermined. So restrictions should be implemented into his logical programming to prevent Capt. Leo from creating his own missions.

Use better monitoring methods

Lack of supervision for a highly classified bleeding edge technological robot was unbelievably stupid. On one hand creating a robot like him could have cost multimillions of dollars, One the other hand if enemies got hold of him, they could have use him to reverse engineer to steal the technology behind him. So it is unbelievable there wasn’t a dedicated team and sophisticated tracking methods to monitor his every step. It’s like military made this expensive and cutting edge robot to just release into the wild to do his own thing only equipped with removable GPS tracker to track his whereabouts.

He is a robot. So his every thought could have been uploaded for monitoring. Use a team or another A.I. to scrutinize those data to detect anomalies and alert if he show any abnormal behavior. Use outside surveillance methods such as CCTV, spy drones, spy satellites etc. to keep an eye out for him apart from his internal tracking methods. Maintain a constant data feed from him to watch what he sees and listen to his conversations so he can’t keep any secrets. If maintaining a constant data feed could compromise his missions or not practical then make him upload those data when he comes back to the base.

Use better hardware

It only took few minutes for Harp to cutout Leo’s tracking and failsafe device. That was a very bad design. There are better DRM methods in consumer software and devices than this one of a kind military robot. So Implement competent tamperproof tracking and failsafe systems and embed it in the system in a way to shut down his systems (brick the robot) if anyone try to forcibly modify it. And better yet implement redundant failsafe systems for assurance. Also do not put such important hardware where they can be accessed using simple tools.

Also Capt. Leo knew how to remove his tracking and failsafe system. A robot does not need to know about its inner workings to do its job. That kind of information should be kept in isolated servers and files with restricted access.

Implement a kill switch

Outside the Wire discusses concerning topics regarding modern and future warfare. One of them is possibility of AI war machines going rogue on their creators. In fact it is a common trope in science fiction genre. Almost all of these stories lack an obvious countermeasure to a AI uprising, a kill switch.

When Lieutenant Harp informed that Captain Leo went rogue, even Colonel Eckhart couldn’t do anything about it. Where was ‘the big red button’ when you needed one? It is a good idea to implement a kill switch for a robot with superhuman abilities just in case if it tries to go all skynet on you. And make sure to make it tamperproof. Then if the robot decided to go rogue, its controlling officers can use the kill switch program to either deactivate it or recall it back to the base or make it self-destruct.

If these kind of safety measures were in place, it could have been difficult for Captain Leo to break free from control of the US military. On the other hand the plot of the movie would have become more complex in order to show him jump through all those safe guards. Actually if the plot had all those protocols and show how Leo overcome all of them to break free, then the movie would have become far more interesting.