The increase in the use of AI in important decision making is in fact something that we should not be excited about. Despite the acceptance of human error and biases, we are turning to alternative solutions what is commonly known as algorithmic bias. Technology is being used to determine who is eligible for mortgages, who gets offered your dream job and who is a potential criminal in your neighbourhood. In general, our ideas about technology that we think are normal are in fact ideas that come from a small homogenous group of people (you guessed it: White Males!). The problem of algorithmic bias is something that simply originated from our history. We don’t know what exact data helped build the algorithm, however racism has been embedded into society over the centuries and now has unconsciously entered the algorithms.
Evidence of this bias became more wide-spread knowledge after the release of the Netflix Documentary ‘Coded Bias’. M.I.T. Media Lab researcher Joy Buolamwini found that the algorithm couldn’t detect her face. Facial recognition software is something rather controversial, despite its increase in adoption. It does not recognize dark-skinned faces accurately, causing unexplainable arrests due to false facial recognition matching leading to discriminatory situations against black people. All of this is happening based on the human prejudices that have creeped into technology. Facial recognition is something that has been described to socially control the public in a way that we don’t see with our own eyes. It is important to make an action on these biases and make legislations before these algorithms are used in society. Without it, no matter what we do to stop racism, it will always be there.
It’s not all bad news though as big tech companies such as Amazon have temporarily stopped providing facial recognition technology to police forces as a commitment to end systemic racism. Hopefully more companies will follow Amazon and join the debate on whether machine learning-based systems should be used in such important decision-making scenarios.
Do you think Amazon and other companies stopping with providing facial recognition technology will halt innovation? A big part of AI is letting it run so it can learn by gathering more information. China has been using facial recognition for a very long time and I feel like, even though there are ethical issues, they will also be the first to benefit from it. Do you think we need to wait for regulations or start with developing the AI and implement it once it’s unbiased? I think we will need more diverse AI developing teams first because the creators influence the technology too much.