Enshrined in instruments such as the Universal Declaration of Human Rights, the International Covenant on Civil and Political Rights, and regional equivalents, human rights and civil liberties represent a foundational framework that protects individuals from abuses of power. 

AI's pervasive data collection capabilities can lead to invasive surveillance, while algorithmic decision-making can perpetuate biases and discrimination if not properly overseen.

The intersection of AI and human rights is broad and multi-directional. AI can be applied in ways that infringe on human rights unintentionally, such as through biased or inaccurate outputs, but can also be intentionally misused to infringe on human rights, such as for mass surveillance and censorship.

Specific rights most commonly threatened include:

• Right to privacy. Through mass data collection, biometric surveillance, and profiling without consent

• Right to equality and non-discrimination. Through biased algorithmic decision-making in policing, hiring, housing, and benefits.

• Right to due process and a fair trial. Through opaque automated decisions that determine guilt, sentencing, or eligibility without meaningful human review.

• Freedom of expression and association. Through chilling effects caused by surveillance of protests, social media, and political activity.

• Right to asylum and freedom of movement. Through AI-assisted border enforcement and migration systems that make life-altering decisions without adequate safeguards.

• Right to life and security. Through autonomous weapons systems and AI-assisted lethal force decisions.

Why it matters

Civil liberties and human rights form the bedrock of democratic societies, ensuring individual autonomy and protecting citizens from the overreach of powerful entities. 

When AI subverts these protections, it shifts the balance of power decisively toward states and massive corporations. 

Without these guardrails, trust in public institutions collapses, social mobility freezes, and the foundational principle that all humans possess inherent, unalienable rights is severely compromised.

Impacts & consequences

When AI systems breach human rights and civil liberties norms, the consequences can be severe and lasting:

• Wrongful deprivations of liberty. Incarceration and loss of agency as a result of misidentification or misuse of facial recognition and other technologies.

• Institutionalised discrimination. AI tools used to screen job applications or assess prospective employees can unfairly discriminate against people of colour, people with disabilities, neurodiverse people, and people from low-income backgrounds.

• Erosion of free expression. Surveillance of political activity and protests suppresses dissent and the exercise of democratic rights.

• Erosion of due process. Unregulated predictive policing and "automation bias" - the dangerous tendency to trust computer conclusions over human judgment - create conditions for mass rights violations.

• Deepened structural inequalities. AI-powered systems embed and amplify historical biases, disproportionately harming already marginalised communities.

• Loss of democratic accountability. When consequential decisions are made by opaque algorithms, citizens lose meaningful avenues for redress.

Causes & contributing factors

Common sources and contributing factors include:

• Biased training data. Systems trained on historically biased data replicate and amplify patterns of inequality and over-policing

• Lack of transparency and explainability. Opaque "black box" models prevent individuals from understanding or challenging decisions that affect them

• Inadequate regulation. Without careful oversight, AI systems used for decision-making have been proven to perpetuate existing systematic inequalities.

• Profit and efficiency incentives. Speed and scale optimisation override careful rights-impact assessment.

• Dual-use design. Tools built for one purpose (e.g. ad targeting) are repurposed for surveillance or enforcement without new safeguards.

• Power asymmetry. States and corporations deploying AI have far greater resources than affected individuals seeking redress. 

• Absent or inconsistent governance. Fragmented, jurisdiction-by-jurisdiction regulation creates gaps that enable abusive deployments.

• Function creep. AI systems originally built for benign or specific purposes (e.g., public health or traffic management) being repurposed for mass surveillance or political policing.