Participating in the Detain/Release simulation was both interesting and emotional for me. It reminded me of the movie Twelve Angry Men. In that film, people must make a fair decision, and they need more than facts. They also need kindness, careful thinking, and the ability to question things (Lumet, 1957).
In the simulation, I did not discuss anything with others. Instead, I had to look at computer-made risk scores. These scores looked scientific and very official. But they were still created by algorithms. O’Neil explains that these systems can repeat unfair patterns from the past even when they look mathematical and neutral (O’Neil, 2016).
As I read each case, I felt pulled in two directions. One direction was the personal stories of the defendants. These stories made me feel empathy and made me want to give them a chance. The other direction was the algorithm’s risk numbers. The Guardian warns that algorithms can “tell lies” when their data is biased or incomplete (O’Neil, 2017). I noticed this when the “high-risk” label made me ignore what the defendant was saying.
Shannon Vallor talks about the “AI mirror.” She says that AI tools copy simple versions of human judgment. They look confident, but they hide their limits (Vallor, 2018). This made sense to me. The colorful bars and charts in the simulation felt very convincing, even when the real situation was more complicated.
I also felt pressure from things like jail capacity and public fear. These forces were not about the defendant, but they still affected my decisions. This showed me that algorithms do not work alone. They are connected to human feelings and rules.
In the end, the simulation taught me something important: AI can help us, but it should not replace human judgment. Fair decisions need both data and human empathy.
References:
Lumet, S. (Director). (1957). Twelve Angry Men [Film]. Orion-Nova Productions.
O’Neil, C. (2016). Weapons of Math Destruction. Talks at Google. [Video]. YouTube.
O’Neil, C. (2017). How can we stop algorithms telling lies? The Guardian.
Vallor, S. (2018). Lessons from the AI Mirror. Santa Clara University. [Video]. YouTube.