
As consumers, we can demand better. Choose tech companies we know support ethical practices and support companies that are aware and are committed to addressing bias in AI.

As consumers, we can demand better. Choose tech companies we know support ethical practices and support companies that are aware and are committed to addressing bias in AI.

What can we do? Advocate for openness and accountability in the design and implementation of FRT systems. Encourage Laws that impose fair data practices and hold businesses accountable for any discriminatory effects.

Governments are intervening as well. A number of communities, including Boston and San Francisco have outlawed the use of FRT by law enforcement until these prejudices are sufficiently addressed. Enacting laws is essential to defending civil rights. nytimes.com/2019/05/14/us/fa…

The leading voices for more equitable FRT are those of women in technology. Scholars, Timnit Gebru and Joy Buolamwini have carried out important research exposing these biases and advocating for moral AI procedures. nytimes.com/2018/06/25/techn…

Tech companies are aware of the issue but process seems to be minimal. Many are currently working to diversify their training datasets and improve algorithmic fairness. Nonetheless, greater accountability and openness are required. media.mit.edu/articles/study…

You may wonder why this matters? Systematic disparities may be strengthened by biased FRT. Marginalized populations suffer the most from the errors in law enforcement organizations use of faulty technology, which exacerbates discrimination and mistrust.

Studies show that FRT systems have shown an error rate of 34.7% for darker skinner women compared to 0.8% for lighter-skinned men. The discrepancy has consequences leading to wrongful arrest and surveillance abuse. #FRT media.mit.edu/articles/study…

Facial recognition technology algorithms are trained on data sets. Lack of diversity in these datasets causes the technology to pick up and reinforce prejudices. Due to datasets being predominantly while male faces, it has lead to higher error for women. nature.com/articles/d41586-0…

Did you know that facial recognition is frequently biased? Facial recognition, used in security, law enforcement, and even in social media, often fails to accurately identify women and people of colour. Here is a thread into why this happens and what can be done: #BiasTech