Data Science & Ethics: Summary
1: Introduction
PRACTICAL INFORMATION
1. Exam (5.6/20)
The exam consists of:
• Large question: Explain and illustrate certain techniques or concepts (/6)
• Discussion case: (/6)
• 5 small questions: Answer in 2 lines (/4)
INTRODUCTION
Data science has a significant impact on:
• Costs and benefits for businesses
• Decisions on humans
• More than just making calls to predefined Python libraries
1. Responsible AI
The double-edged sword of advancement
It was the best of times... It was the worst of times...
• Reduce risk (e.g., predict terrorist attacks) • Data leaks
• Reduce crime (e.g., detect tax fraud) • Discrimination*
• Increase profitability • Digital pawns
,It was the best of times... It was the worst of times...
• Improve medical diagnosis • Filter bubble
• Increased "good" • Increased "bad"
*Models are built to discriminate good from bad, but it shouldn't do so based on
ethnicity or other factors.
Responsible AI = knowing what is right and wrong when working with AI.
1.1 Data science and ethics
• Ethics = moral principles that control or influence a person’s behavior or the
conducting of an activity.
• Moral = connected with principles of right and wrong behavior.
o Law = what you can do
o Ethics = what you should do
Data science ethics = the domain of what is right and wrong when doing data
science.
Responsible AI = the development and application of AI that is aligned with the
moral values of society.
2. Why care?
① Expected from society
We have reached an age where society expects business leaders and data
scientists alike to be ethically responsible.
Generation Z
• Born in 1995 - 2010
• 90 million in the US
• Cares about social justice and ethics
,② Huge potential risks
Be aware of the risks and countermeasures!
Risks for humans Risks for business
• Physical and mental well-being (self-driving)
• Reputational
• Privacy
• Financial
• Discrimination
AI in banks determines whether you get a loan or not, which has a big influence
on your life. It is important to know how these AI systems work and whether all
ethical steps are considered.
③ Potential benefits
Understanding ethical concerns and applying techniques to deal with this can
improve the data model and be a marketing instrument.
• Remove bias in data: Improve the accuracy and fairness of the model
• Explain predictions: Improve the trust in the model
• Ensure proper data gathering: Better data quality
• Part of a company’s brand
④ The AI Act
• Risk-based approach to regulate AI
• Applies for every company that works with EU citizens
Requirements for high-risk uses of AI:
• Establish safeguard against biases in datasets
• Prescribed data governance and management practices
• Ability to verify and trace back outputs through the system’s life cycle
, • Including provisions for acceptable levels of transparency and
understandability for users of the systems
• Appropriate human oversight
A risk-based approach:
• Unacceptable risk (e.g., AI systems that threaten the safety of people):
banned
• High risk (e.g., AI systems used in employment, law enforcement or
migration): strict obligations
• Limited risk (e.g., chatbots)
• Minimal risk (e.g., spam filters)
2.1 Summary
There are multiple reasons to care about data science ethics:
• Life goal in itself (philosophical goal)
• Societal and business reasons:
o Expected from society
o Huge potential risks
o Data science ethics can bring value
o AI act
MACHINE LEARNING
Note: Data scientists and business students are not inherently unethical, but at
the same time not trained to think this through neither.
Machine learning = automatic extraction of knowledge from data.
→ Prediction of new things based on patterns detected in historical data.
1: Introduction
PRACTICAL INFORMATION
1. Exam (5.6/20)
The exam consists of:
• Large question: Explain and illustrate certain techniques or concepts (/6)
• Discussion case: (/6)
• 5 small questions: Answer in 2 lines (/4)
INTRODUCTION
Data science has a significant impact on:
• Costs and benefits for businesses
• Decisions on humans
• More than just making calls to predefined Python libraries
1. Responsible AI
The double-edged sword of advancement
It was the best of times... It was the worst of times...
• Reduce risk (e.g., predict terrorist attacks) • Data leaks
• Reduce crime (e.g., detect tax fraud) • Discrimination*
• Increase profitability • Digital pawns
,It was the best of times... It was the worst of times...
• Improve medical diagnosis • Filter bubble
• Increased "good" • Increased "bad"
*Models are built to discriminate good from bad, but it shouldn't do so based on
ethnicity or other factors.
Responsible AI = knowing what is right and wrong when working with AI.
1.1 Data science and ethics
• Ethics = moral principles that control or influence a person’s behavior or the
conducting of an activity.
• Moral = connected with principles of right and wrong behavior.
o Law = what you can do
o Ethics = what you should do
Data science ethics = the domain of what is right and wrong when doing data
science.
Responsible AI = the development and application of AI that is aligned with the
moral values of society.
2. Why care?
① Expected from society
We have reached an age where society expects business leaders and data
scientists alike to be ethically responsible.
Generation Z
• Born in 1995 - 2010
• 90 million in the US
• Cares about social justice and ethics
,② Huge potential risks
Be aware of the risks and countermeasures!
Risks for humans Risks for business
• Physical and mental well-being (self-driving)
• Reputational
• Privacy
• Financial
• Discrimination
AI in banks determines whether you get a loan or not, which has a big influence
on your life. It is important to know how these AI systems work and whether all
ethical steps are considered.
③ Potential benefits
Understanding ethical concerns and applying techniques to deal with this can
improve the data model and be a marketing instrument.
• Remove bias in data: Improve the accuracy and fairness of the model
• Explain predictions: Improve the trust in the model
• Ensure proper data gathering: Better data quality
• Part of a company’s brand
④ The AI Act
• Risk-based approach to regulate AI
• Applies for every company that works with EU citizens
Requirements for high-risk uses of AI:
• Establish safeguard against biases in datasets
• Prescribed data governance and management practices
• Ability to verify and trace back outputs through the system’s life cycle
, • Including provisions for acceptable levels of transparency and
understandability for users of the systems
• Appropriate human oversight
A risk-based approach:
• Unacceptable risk (e.g., AI systems that threaten the safety of people):
banned
• High risk (e.g., AI systems used in employment, law enforcement or
migration): strict obligations
• Limited risk (e.g., chatbots)
• Minimal risk (e.g., spam filters)
2.1 Summary
There are multiple reasons to care about data science ethics:
• Life goal in itself (philosophical goal)
• Societal and business reasons:
o Expected from society
o Huge potential risks
o Data science ethics can bring value
o AI act
MACHINE LEARNING
Note: Data scientists and business students are not inherently unethical, but at
the same time not trained to think this through neither.
Machine learning = automatic extraction of knowledge from data.
→ Prediction of new things based on patterns detected in historical data.
