The face of a riot police officer captured over 180º from footage of a Kill the Bill protest on 17/04/21, isolated and enhanced as a manual replication of the computer vision processes used by facial recognition software; reversing the surveillance gaze against the face of the state.

RAKE Collective, consisting of Nancy Hurman, Flora Thomas, Vera Zurbrügg and Nuno Guerreiro de Sousa, is a visual research collective using open source data to investigate a variety of unseen elements in society, business and politics. The themes explored in their collaborative work include human rights violations, government censorship, surveillance and bureaucratic violence.

 interrogates the systems of control put in place by the UK government in the name of order and security. Catalysed by increasing police powers and ongoing bureaucratic violence, the work examines surveillance and technology to visualise the ways in which the state polices its people.

The global number of CCTV cameras is set to reach one billion by the end of 2021, while facial recognition and biometric systems are being integrated into everyday life and deployed as tools of state control. Initially focusing on London, the third most surveilled city in the world, the ongoing project explores how the ubiquitous and ever-growing network of surveillance is implemented, and the multiple points at which this technology intersects with – and influences – humanity.

The work invites the viewer to return the glare of the state, looking past the camera to challenge the systems of power which govern our everyday lives.

Images created using a generative adversarial network (GAN), pitting two AI against one another and tasking one to produce outputs that the other would recognise as ‘real’. The model has been trained on a dataset of images of CCTV and surveillance cameras scraped from the internet, generating images which represent the algorithm's own vision. The algorithm can be seen to ‘fixate’ on certain details that might be missed by the human eye, and within this process these data points become disproportionately visible. These abstract results illustrate the disparity between AI’s comprehension and our own.

The UK’s hurried push-through of the Police, Crime, Sentencing and Courts Bill would give the police and the Home Secretary sweeping new powers to restrict protest, expand already 

 Traveller communities. The bill would also empower a system of police-led data gathering, storage and sharing; leading to increased predictive policing - the use of which has seen technology implemented with little oversight, operating under a false sense of machine neutrality while continuing to perpetuate bias and discrimination.

Who has the authority to look? In the context of policing, surveillance and 

, the right to look - to surveil - becomes an apparatus of power. By manually reproducing processes of image analysis, facial recognition and algorithmic generation against agents of the state - both man and machine - and engaging with open source material and citizen media, Police State aims to challenge the surveillance gaze; highlighting who these systems benefit and who they seek to control.

Stills and excerpts from protest videos – gathered through activist networks, combed from social media and collected by RAKE members – are laid over Google Street View locations where the events took place. The clash between generic street views and grainy citizen media highlights the tension in public space that is increasingly surveilled and controlled.