Body cameras don't resolve police-civilian disputes because viewers' group identities determine what they see. Jurors identifying with police focus on the suspect's faults, while others focus on the officer's, leading to opposite conclusions from the same footage.

Risk assessment tools used in courts are often trained on old data and fail to account for societal shifts in crime and policing, creating "cohort bias." This leads to massive overpredictions of an individual's likelihood to commit a crime, resulting in harsher, unjust sentences.

Ben Horowitz reveals that a major source of violent police encounters stems from inaccurate suspect descriptions. By funding the Las Vegas PD with AI cameras, they can identify the correct vehicle or individual with certainty, preventing dangerous confrontations with innocent citizens and enabling safer apprehensions.

Work on this topic must be careful to avoid inflammatory framing. A fiery, un-nuanced approach risks politicizing the issue, making it harder to build the broad coalitions necessary for effective action. The goal is to solve the problem, not to create ideological battlegrounds.

When AI systems are trained on historical data, such as past hiring or policing records, they learn and perpetuate existing societal biases. This creates a dangerous illusion of objectivity, where discriminatory outcomes are presented as neutral, data-driven "predictions" by an algorithm.

When communities object to surveillance technology, the stated concern is often privacy. However, the root cause is usually a fundamental lack of trust in the local police department. The technology simply highlights this pre-existing trust deficit, making it a social issue, not a technical one.

While AI can inherit biases from training data, those datasets can be audited, benchmarked, and corrected. In contrast, uncovering and remedying the complex cognitive biases of a human judge is far more difficult and less systematic, making algorithmic fairness a potentially more solvable problem.

The movement to defund the police doesn't eliminate the need for security; it just shifts the burden. Wealthy individuals and communities hire private security, while poorer communities, who are the primary victims of crime, are left with diminished public protection.

As the pace of AI-driven change and information generation accelerates, actors like journalists and courts may be unable to keep up without using AI assistants. This creates a dangerous dependency, forcing them to rely on potentially biased systems controlled by the powerful entities they are supposed to hold accountable.