Considering how things are moving so far, it is safe to say that the time has already come to face the truth that AI based police interrogation will really be a thing. When it comes to law and order, it might function slowly, but normally police or army interrogations follow no protocol and resort to every possible way to break a detainee. Most people might think what harm a chat box can really do, physically nothing I guess, but for a detainee who will be interrogated for 7 consecutive days or more, the AI is not limited to a 5minute response and it can go on interrogating for an infinite period till the detainee breaks. It is the basic practice to break a detainee into confessing any crime through interrogation as it is the first protocol. In this case, only getting the confession of the detainee is not the issue of any interrogation, but getting tangible proof is. So, following that system making the interrogation intense can never be any problem as without proof no law enforcer can submit their evidence in court supporting the government and fight against the detainee. This is the basic reason every interrogation is made intense willingly, so that the law enforcer can get tangible evidence against the detainee fast and quick. However, no AI is made by any alien, but human and so the interrogating AI’s question will be the same as any regular police that you see in the movies. However, following the movie version doesn’t always turn out right and it should be as leading and accusing as anyone can imagine with every possible scary sound effects to go with it. Since it has not been used in reality, it is safe to have a good imagination in portraying the effect of such an AI. Since, nobody is actually harming the person doesn’t really make the issue safe but mentally more challenging as physical pain has a chance to go away or at lessen after passing of certain time. But, an interrogation protocol, where the confirmation needs to be achieved from the detainee through sound and fear, if such an interrogation takes place for 7 consecutive days, it is bound to make the detainee confess into anything or have a heart attack or be a mental patient without any doubt. Normally, according to the movies, the interrogation is always done in a dark filthy room beating the detainee into confession. So, in order to create an environment of fear it is bound to make the interrogation in a dim or dark filthy room which is I guess bound to happen as the source of the voice needs to be kept hidden. Now, when it comes to AIs, think about all the apps out there and their capabilities, one can correct speeches, one can change voices, one can forge images, and some even can create forged realities or videos? How do you think in a world filled with such diverse AI applications, if this very thing can be applied on detainees in the name of eliminating physical abuse from the justice system would turn out to be. It will be something more harmful than anything anyone can ever dreamt of as the purpose of any interrogation is to get the proof from the detainee by any means possible within a very short time. First, to create the environment of fear, the conversation will take place ensuring various sound effects are associated with it to create fear. Then, each and every conversation will start leading with accusations and make sure the answers of the detainee can be corrected or accusations be corrected according to the replies of the detainee leading to a more fear inflicted environment as we already know like any normal human if an AI uses loopholes in conversations where the issue is not to proof the AI is right, adding more allegations won’t be a problem. However, we are always under the impression that all AIs follow moral or ethical boundaries, but all the functions the current AIs are doing like correcting speeches, changing voices, changing images or changing voices or even creating forged videos are actually not considered as ethically moral. So, now think of the situation from a government’s perspective as they need to prove the detainee convicted of the crime and submit the proof to the court within the allotted 7days and to ensure that they will need everything in their arsenal to prove according to the given allegation. Unfortunately, in such a time bound protocol, they will resort to all the advantages that have been offered by the AI as being part of the legal system they are authorized to perform all tasks necessary to get to the truth without harming the detainee in a visible manner. So, the scary part will be when an AI changes the voice into one of the known supreme characters of the government and next the person’s most feared character having conversations with the person’s closed ones threatening them in every possible way with a sound effect of a crowd acknowledging the claims to intensify the issue further. If the detainee breaks it's good and if not, two or three forged images or even videos won’t be any problem to support the governmental claims of allegations won’t hurt to break the detainee into submission as time allotted by the court is very limited. So, see, considering what the current AI can do, framing anyone into any possible allegation won’t be any problem with proper tactics used to allege him with precise timing. As the responses of the AI are prepared at light speed, the detainees next responses could also be analyzed to predict his responses by saying simple words like is this what you want to say or changing his conversation midway to portray his predicted responses aligned with the allegation. In a fearful environment, the detainee would appear as if he is in a voodoo world, where he is against not only a person but the entire system and it is scaring him from top to bottom with all his responses predicted. It is bound to make the person break at any possible time or frame any person with any allegations.
