Faception

Transparency and accountability 🙈

Faception is seen to suffer from several important transparency and accountability issues: 

• Algorithmic decision-making. Faception's algorithm is proprietary, with the company refusing to disclose the methods used to analyse facial features and make predictions.

• Privacy. Faception collects and analyses sensitive biometric data, including facial images. However, the company's data protection policies and procedures are not publicly accessible or transparent, and it is unclear how this sensitive information is stored, shared or protected.

• Deployment. Faception refuses to say how its technology is used in various contexts, including law enforcement, border control and employment screening, making it difficult to assess the potential risks and benefits of its technology.

• Errors or misclassifications. Faception's technology is liable to errors or misclassifications, but the company has failed to establish clear procedures for addressing these kinds of errors or to provide recourse for those wrongly affected.

• Regulation and oversight. Faception operates in a regulatory grey area, and there is limited oversight of its activities. This lack of regulation can lead to unchecked use of the technology, potentially resulting in harm to individuals or groups.

Risks and harms 🛑

Faception's claims to predict complex personality traits and behaviours such as identifying terrorists or paedophiles has been criticised for potentially leading to false positives and unjust profiling.

Specifically, the company's claims about the accuracy of its technology are based on largely unproven scientific theories about the relationship between facial features and personality traits. 

Furthermore, its technology appears to be based on Western cultural and social norms, which may not be applicable in other cultural contexts, and it seems not have have demonstrated an understanding of the potential cultural and social implications of its technology.

Research, advocacy 👩🏼‍⚖️

• Bucher T. (2022). Facing AI: conceptualizing ‘fAIce communication’ as the modus operandi of facial recognition systems