The rise of mass audio surveillance
A number of countries - and we shall not provide even a partial list here - have begun selective procedures of mass audio surveillance of their citizens. In this article we explain what this involves, its consequences for both the individual citizen and for society as a whole, and we form an assessment of whether this is in the greater good and whether international legal measures ought to be taken to suppress it.
Mass audio surveillance is a phenomenon that has become possible with the exponential increase in processing power of modern computers. It involves using various electronic devices in public places - including people's mobile telephones, that have built-in microphones; and CCTV, which can lipread the people they observe (even private CCTV footage is available to governments, even when the private owners of the CCTV think they have turned the CCTV off - things not commonly understood) - to collect recordings of what people are saying.
Government IT systems then create machine transcripts of what is being said using computer algorithms that at least in part deploy artificial intelligence technology to interpret the sounds they hear into words. Certain words are then recognised as trigger words by the computer system running the programme of mass audio surveillance, and whenever a trigger word is heard by the system the matter is referred to a human being (typically a member of the country's security or intelligence services) for some sort of security-based review.
In principle, such a system may seem entirely laudable. It involves the use of technological advances to render the fighting of crime more efficient and effective. Just as has happened for a number of years with emails and other electronic communications, audio communications are scanned for certain words that may be associated with criminal activity. Thereby actual or potential criminals may be identified and selected for closer scrutiny for the Police and internal security authorities. Civil liberties concerns may arise if material of this kind is admissible as evidence in court without a prior lawful judicial warrant; but with a tolerably robust legal system and exemplary standards of ethics and professionalism amongst the Police and internal security authorities a development of this kind, using modern voice recognition software, artificial intelligence and computational capacity to fight crime may be argued to be a thoroughly good thing.
The problems arise where the aforementioned standards of exemplary government conduct are not met. The legal system of the country in question may be less than robust. It may not harbour rigorous rules about the admissibility of evidence acquired by such means. Or the prosecutors and courts may privately admit evidence that publicly they know they are not allowed to do, denying defendants due process. In other words, Judges and Prosecutors may get whispers of varying kinds from members of the law enforcement authorities suggesting what inadmissible evidence says and inviting them to skew their legal judgments accordingly.
Alternatively - and every bit as concerning - is the situation where the key words used are nothing to do with legitimate law enforcement at all, but are chosen to identify persons the government of the day may consider as subversive. Being subversive (that is to say, being opposed to the government of the day) is or should not be a crime, unless one's subversion involves the use of unlawful force with a view to overthrowing a legitimate governmental regime. Nevertheless some countries with less than perfect democratic traditions have historical habits of collecting wanton data about individuals with a view to profiling them for their legitimate political opinions and legitimate political activities. This is not something governments should properly do, and the danger associated with contemporary mass audio surveillance is that governments use it for improper purposes and then harass the people they have identified as subversive or otherwise legitimately opposed to the government of the day.
The gravest danger with mass audio surveillance is that it facilitates a sort of insidious modern totalitarianism, in which people learn that what they say is being listened to by a computer and they even learn the key words that trigger review of what they have been saying, because they have friends or contacts in the police or security forces. In societies in which membership of the security forces is ubiquitous, this is virtually inevitable. (In some societies, ordinary people are forced to act as de facto agents of the internal security forces - in essence, informants - often for free.)
Therefore people start to engage in self-censorship over subjects that are really harmless to talk about and do not indicate any sort of wrongdoing. In self-conditioning the way they speak, they actually start to self-condition what they think about; this was the result of Stalinism in the Soviet Union and it changed the way people think about their daily lives. It contributed towards the creation of what we have described as the "not" method of communication, in which you say the opposite of what you really mean in order to evade the attentions of the security forces. Mass audio surveillance, like any system that regulates what people talk about, is a form of censorship. It detracts from common sense and causes people to talk in euphemisms and deliberately to avoid certain constructions, and instead to talk in silly ciphers to avoid the sanction of the censor.
The net consequences for society of mass audio surveillance is regulation of free speech. In the West we consider this unabashedly a bad idea, because we believe that it is only through the open and free expression of views and their debate that the truth prevails, as do superior ideas. By stopping us from talking about certain things, we prevent that process of free debate that is a cornerstone of western civilisation going back as far as the Enlightenment. In the words of Winston Churchill, “I disapprove of what you say, but I will defend to the death your right to say it”. This principle is eviscerated by mass audio surveillance.
What should we do about this phenomenon, in which it is possible to regulate free speech through increases in information technology by having machines listen to everything we say and parse our language so as to prevent us from saying things - thereby producing a chilling effect upon freedom of speech? The answer is that we must ban it unless it takes place a judicially authorised warrant, just as other sorts of intrusion into our private lives and civil liberties are banned without warrants showing probable cause (the test in US legal language) or reasonable suspicion (the equivalent test in British English legal language) that the parties to be placed under illicit surveillance are genuinely engaged in some serious criminal wrong. Illicit intrusion upon people’s conversations, whether it be through a traditional wiretap, hacking someone’s email or electronic audio surveillance, is a gross infringement of civil liberties and can be justified only by an independent judicial authority.
In this regard the West, leading the world in the advancement of civil liberties, must take a principled stand and should enact legislation explicitly outlawing this type of surveillance whether it be undertaken privately or by a public authority. We cannot hope to win the war of ideas with the world’s undemocratic regimes, and maintain the moral authority of western ideals, without taking a lead upon this new and unnerving phenomenon. Explicit criminalisation of mass audio surveillance, absent judicial warrant (that in due course the subject of such a warrant should have the opportunity to challenge in a legal procedure exhibiting transparent due process), is an essential step to preserve the western value of liberty. Democracy cannot survive without liberty, because a free vote becomes meaningless if there cannot be debate about the ideas that the competing candidates for one's vote are advocating. Nothing less than the Western way of life is at stake.