How police manipulate facial recognition

Summary

This video explores the controversial use of facial recognition technology by police, highlighting a bizarre case where the NYPD used a photo of actor Woody Harrelson to catch a shoplifter. The video critiques how facial recognition systems can be manipulated and misused, emphasizing the risk of false matches leading to unnecessary police stops. It also addresses the accuracy issues, particularly for women and people of color, and the lack of proper oversight in the use of such technology. While some argue it's an effective tool for law enforcement, others, like San Francisco, have banned it due to concerns about privacy and abuse.

Takeaways

• 😀 Facial recognition technology was used by the NYPD to catch a shoplifter, despite lacking a clear image of the suspect.
• 😀 The NYPD used a photo of actor Woody Harrelson, based on the clerk’s description, and it led to a match with the actual suspect.
• 😀 The use of facial recognition in law enforcement raises ethical concerns about how it should be applied and whether it should be used at all.
• 😀 Georgetown University conducted a study revealing that some police departments manipulate facial recognition images to force a match.
• 😀 Facial recognition works by tracking unique facial features such as pupil distance and cheekbone shape, but accuracy is best with clear, straight-on photos.
• 😀 Technology providers like Amazon and Google have made facial recognition more accessible, enabling almost anyone to build such systems.
• 😀 The accuracy of facial recognition is not standardized, meaning police can adjust the system’s thresholds to achieve a match, even if the match is weak.
• 😀 Facial recognition systems are less accurate for women and people of color, leading to higher error rates for these groups.
• 😀 While the NYPD claims facial recognition isn’t solely responsible for arrests, it has been involved in over 2,800 arrests over five years.
• 😀 Some cities, such as San Francisco, have banned facial recognition for police use, citing concerns over lack of oversight and potential misuse.
• 😀 The controversy surrounding facial recognition highlights broader issues with surveillance technology and the need for proper oversight to prevent misuse.

Q & A

How did the NYPD use facial recognition to catch a shoplifter, and what was unusual about it?

The NYPD used facial recognition to catch a shoplifter who was identified by a clerk as resembling actor Woody Harrelson. They uploaded a photo of Harrelson to the facial recognition system, which worked and led to the arrest. This was unusual because the system was never designed for this purpose, and the identification was based on a resemblance rather than a clear photo of the suspect.

What issue does the Georgetown University study highlight about the use of facial recognition by police departments?

The Georgetown University study revealed that some police departments have manipulated facial recognition systems by pasting in missing facial features (like a blocked eye or an open mouth) to get a match. This shows how police can alter the input to achieve a desired outcome, which raises concerns about the reliability and ethical use of these systems.

How does facial recognition technology work in general?

Facial recognition technology works by tracking key facial landmarks, such as the distance between pupils, the angle of the nose, and the shape of the cheekbones. These features are distinctive and can be tracked across photos, even at angles or with parts of the face blocked. The system uses algorithms to match these features across different images.

Why is facial recognition more accurate for certain demographics, and how does this affect its use by the police?

Facial recognition systems are generally less accurate for women and people of color, which is likely due to the algorithms being primarily trained on images of white men. As a result, these groups are more likely to experience false matches, which increases the risk of unwarranted police stops and potential violations of civil rights.

What are the risks of using facial recognition in policing, according to the transcript?

The main risks of using facial recognition in policing include false matches leading to unwarranted police stops, the potential for racial bias due to inaccuracies in the system, and the lack of clear oversight in how these systems are implemented. Additionally, without strict rules about how the technology is used, it could lead to a violation of individuals' rights.

How did the NYPD's use of facial recognition contribute to the number of arrests, even if no arrests were made based solely on the technology?

Although the NYPD states that no one has been arrested solely on the basis of facial recognition, the technology has been involved in more than 2,800 arrests over the five and a half years the program has been running. False matches or potential matches can still lead to police stops and investigations, even if no arrest results.

What is the controversy surrounding Detroit's Project Greenlight, which uses facial recognition?

Detroit's Project Greenlight involves a network of surveillance cameras with facial recognition technology, credited with a 23% drop in crime. However, some community members oppose it due to concerns about lack of transparent oversight and the overwhelming number of tips that flood the police force, raising questions about its effectiveness and the balance of privacy versus security.

What are the concerns that led to the ban of facial recognition by police in San Francisco?

San Francisco banned the use of facial recognition by police due to concerns about privacy, lack of oversight, and potential abuse of the technology. Critics, including Supervisor Aaron Peskin, argued that facial recognition technology represents 'Big Brother' surveillance and that the government cannot be trusted to handle it properly.

What is the core argument against facial recognition technology in the context of government use?

The core argument against government use of facial recognition technology is that it lacks sufficient oversight and transparency. Critics argue that the risks of abuse, errors, and violations of privacy outweigh the potential benefits, especially when the technology can lead to unjustified police stops or arrests.

How have the large tech companies like Amazon and Google impacted the accessibility of facial recognition technology?

Amazon and Google have made facial recognition technology more accessible by integrating it into their cloud computing services. This allows almost anyone with basic coding skills to create their own facial recognition system for a relatively low cost, raising concerns about the widespread availability of such powerful surveillance tools.