The Tell-Tale Code: AI in Global Surveillance Systems, Privacy and the Democratic State.

His eye was like the eye of a vulture, the eye of one of those

terrible birds that watch and wait while an animal dies, and then fall

upon the dead body and pull it to pieces to eat it.

Poe. The Tell-Tale Heart.

In our contemporary age, there are eyes everywhere. Surveillance is no longer a public action, monopolised by certain administrative institutions as an extension of the RSA systems governing nation-states, but a private maneuver as well. Technology, especially Artificial Intelligence has aided the development of our internal and external security systems, in aspects that were otherwise impermeable by man-made interventions. In an era where concerns of privacy and confidentiality, information security and disinformation are forefronted on account of the technological revolution, one must question their monitored existence, especially as citizens of an automated democracy.

AI-Assisted Surveillance Systems

As most AI-assisted software are concerned, surveillance and security systems too operate on the basis of Machine Learning and computer vision. AI video surveillance analysis involves a more complex method of motion detection which delves into the classification of objects on the basis of their movement and behavior. Having been trained on large datasets of audio-visual input, these systems are able to detect anomalies through comprehensive pattern detection and predictive analyses based on repetitive behaviours. Repetition training becomes crucial in the conception of these systems to counteract upon possible security breaches, like theft, trespassing or hacking. Facial recognition, likewise, is another imperative aspect offered by AI surveillance systems. For most people, facial recognition is limited to unlocking their mobile devices (Which too is a form of public surveillance), however, government institutions, as well as travel institutions such as Airports, have opted for facial recognition kiosks as the first stage of identification in official matters to prevent impersonation and fraud.

Modern surveillance systems also possess the ability to instigate alarms or 9-1-1 emergency calls in the case of any detected disasters such as fires or break-ins. Additionally, monitoring systems employed in child-care or elderly-care facilities are inconceivable without AI intervention as such instances require an immediate transfer of information pertaining to the wellbeing of the individuals being monitored. According to a study by Deloitte, the possibility of crime reduction initiated through AI surveillance systems is around 30-40 percent, while emergency response times have been lowered by 20-35 percent. Public surveillance requires biometrics, drones, CCTV, black boxes, etc, to ensure public safety. In other words, the modern surveillance system cannot exist in antiquity without AI detection and analyses. 

For those who are an avid watcher of Shark Tank might be familiar with Jamie Siminoff’s Ring, which was later acquired by Amazon. Ring, like Qubo, is not merely a video doorbell, but also a surveillance system. We will delve into the intricacies of private AI surveillance systems assisting state surveillance later in this blog; However, the point remains that otherwise mundane security systems, when technologically-enhanced through the incorporation of AI, become surveillance systems. 

Amazong Ring

The Unforeseen, The Unforecasted

As far as most AI-assisted tools are concerned, privacy and bias remain at the forefront of all discussions pertaining to the topic. In terms of AI-assisted surveillance, bias is undoubtedly one of the more pressing issues that could possibly affect a plethora of people. Having been trained on highly specific datasets, dynamically growing as new information is processed, AI-surveillance systems too may suffer from AI-induced biases as these datasets could be inherently biased. There are chances where these systems can discriminately profile the people and objects they are expected to analyse. Additionally, there are security systems that particularly target a specific section of the population, mostly used for monitoring regions with a higher population of immigrant or displaced individuals. 

Besides biases, another potential risk is that of cybersecurity or data leaks. Private surveillance systems are expected to retain sensitive and private information within the boundaries of the owners or other authorised personnel (System-developers), however, these household systems tend to become the rudimentary hotspot for public surveillance, delving into the disturbing question of ‘Who is really watching us?’ As mentioned earlier in this blog, Ring, the Amazon-owned video doorbell system, is abundant in households across the Global West, especially in the United States, similarly to Qubo’s popularity in the Global East. Earlier this month, during the protests over President Trump’s crackdown of immigrants in LA, LAPD officers had revealed that they were using facial recognition technology to identify and locate immigrants protesting ICE raids. While Amazon had earlier discontinued a feature that enabled law enforcement to request information from the Amazon Ring Neighbours platform, federal and local cops could still acquire data through court orders. On the flipside, civilians have been using neighbours to spread information on ICE raids to warn potential immigrants. 

From a top-bottom approach, private security systems provide the perfect avenue for both law enforcement and administration to function “effectively”, regardless of the individual rights of the citizens being jeopardized. This also indicates the possibility of cybersecurity attacks, wherein large amounts of individual data could be sold to various repressive institutions. At this point, without proper safeguards, firewalls or transparency in policies and data-usage, the whereabouts of our data is rather dubious.

Ring Neighbours

Surveillance or Subjugation?

Authoritarianism, Totalitarianism, AI-tarianism- What are we headed towards?

To understand the current state of public surveillance, it would be necessary to understand the global technological market and its key actors in regards to the production of these systems. For starters, 75 out of 176 nations currently rely upon AI for surveillance purposes. China is the major producer and exporter of AI surveillance systems. With corporations such as Huawei and Dahua, they currently supply these systems to over 63 countries, wherein 36 of them have ratified China’s Belt and Road Initiative (BRI). For nations such as Laos, Uzbekistan, Kenya, Mongolia, etc, China, in addition to the surveillance systems, also provides soft loans to encourage additional purchases. The AI Surveillance industry seems to be a duopoly more than anything with both China and the United States at its forefront. More than half of the existing surveillance cameras worldwide are located within China, and they were revealed to have been installed for the purpose of “controlling and managing people.” Besides covert surveillance systems, China has also introduced robotic police patrol agents to patrol public areas which record civilians in search for individuals who are wanted by law enforcement. It is truly a modern-day panopticon.

The biggest issue that these surveillance systems have brought forth would undoubtedly be genocide or the oppression of minority communities. For instance, the Uyghur Muslims, a Turkish ethnic group, in the Xinjiang province of China have been tracked in the province, which is generally regarded as the epicentre of surveillance technology in China,through digital checkpoints, facial recognition and mandated “health checks”. An individual once detected to be a part of this ethnic minority triggers an “Uyghur Alarm” to alert nearby civilians. No one is truly secure anywhere. 

AI Surveillance ensures that the power which was distributed across various individuals, is concentrated with fewer people. In other words, AI is also taking over occupations in this section, but as far as law enforcement is concerned, it becomes even more dangerous considering the fact that this would inevitably result in jurisdiction monopolies and authoritative control. AI Surveillance ensures 24/7 monitoring capabilities, which when put into perspective, means that one is always under supervision. While one would assume that it is mostly authoritarian nations employing these technologies, that could simply be farther from the truth. 51% of advanced liberal democracies employ AI surveillance systems. However, this does not necessarily mean that these systems are being abused in such instances, but the possibility still remains, as well as a gradual shift towards it.

Conclusion

All in all, the panopticon is nigh. Right now, democracies must ensure the development of clear, ethical frameworks to limit how these systems are utilised. What civilians require the most is transparency and privacy protections. Without considering the socio-cultural and political implications of the technologies being implemented, it is not possible for a liberal democracy to function, and such an aspect is one which policymakers should treat as of paramount importance. Otherwise, we are undoubtedly headed towards an international state where civil liberties cease to exist and one is under constant surveillance. 

In other words, it is the dystopian future that we dread. Where the Big Brother is always watching.