Childcare benefit scandal was a problem that happened in The Netherlands. Between 2005 and 2019, authorities wrongly accused an estimated 26,000 parents of making fraudulent benefit claims, requiring them to pay back the allowances they had received in their entirety. In many cases, this sum amounted to tens of thousands of euros, driving families into severe financial hardship.
This was, in part, caused by an AI system, that discriminated people.
These AI algorithms are used to discriminate (meaning: separate in different categories):
However, in many cases this leads to problems, as depicted above.
Bias is kinda necessary to obtain the desired results (distinguish A from B), but can also lead into problems related to sexism and racism.
Categorization can be used to influence people, but it can also lead to unintentional influences.
In CI, you have learned some specific techniques:
When we apply them to specific problems, we need to think about a lot of other things…
Let’s train a very simple neural network:
We train it with the following pictures: