Coded Bias starts with explaining how MIT Media Lab researcher Joy Buolamwini first started researching into bias which is built into Artificial Intelligence (AI) based advanced technology as a result of the use of biased datasets. The documentary, which primarily tackles facial recognition technology, aims to highlight how people embed their own unconscious biases into AI.
The documentary follows the journeys of multiple women as they try to sound the alarm against the use of this biased AI. They warn against not just inaccurate facial recognition systems which fail to correctly identify certain groups more than others, but also the use of these technologies by governments moving towards authoritarianism. UK based Silkie Carlo, a director at the non-profit organisation Big Brother Watch, talks about how she is worried that this is “China-style surveillance for the first time in London”. Kantayya then moves from state sponsored mass surveillance to corporate surveillance. Data, which we give up without a second thought, is being aggregated by a small group of companies who are building the future of AI.
One of the points that the documentary makes is that AI replicates the world as it exists. Thus, since we do not live in a perfect world, how can the AI be perfect? The decision making that AI performs is mathematical and not ethical thereby failing the need for making social progress.
The brilliance of the documentary lies specifically in the perspectives that it presents. While Social Dilemma highlighted how someone who built this technology belatedly realised how harmful it is, Coded Bias highlights the perspectives of researchers who were against the unfettered use of AI since day one. Thus, it manages to convey its point without feeling the need for unnecessary dramatisation. However, if you do want to look at the worse case scenario, I suggest you watch Tom Cruise’s Minority Report next.