This piece was created for my friend's English presentation, constituting 20% of the final grade.
Our theme revolves around "Something amazing in sciences". The group has chosen the topic of ethical challenges arise with AI in the job market.
4. BIAS AI TOOLS
Situation in which a machine learning system
discriminates against a particular group of people
Generally caused by data scientists who collect the
information
MISSING POSSIBLE ASPECTS
AND CASES ?
Trained with incomplete data
Naturally make mistakes
POSSIBLY LEAD TO DISCRIMINATORY DECISIONS,
FAVORING OR PENALIZING CERTAIN GROUPS
UNFAIRLY
5. AUTOMATED
RECRUITMENT SYSTEM
Developped by Amazon in 2014
Supposed to select the best CVs among candidates for
an open position
HOWEVER
Proven to discriminate against women
WHY ?
Model trained on historical data from Amazon employees
which the vast majority of them were white men
6. FAIRNESS
Automated recruitment system that gives disproportionate
weight to specific skills without considering cultural or
socio-economic contexts
FOR EXAMPLE,
Essential aspect in the use of AI in the labor market
discriminates against a particular group of people
Ensure that automated processes do not create unfair
advantages or disadvantages for certain groups
MAJOR ETHICAL CONCERN
8. ETHICAL GUIDELINES
Identify potential sources of bias
Explains to users and candidates how decisions are
made
Promoting trust in the system
TRANSPARENCY AND EXPLAINABILITY
Representative and diverse datasets
Reduce the risk of bias
DATA DIVERSITY