Intro to Data & AI Ethics

Define ethics as moral principles guiding what is right or wrong, and examine data ethics for data collection, sharing, and use, alongside AI ethics for development and deployment.

Explore the Cambridge Analytica scandal, where a personality quiz harvested user data to target political ads and misinformation, prompting privacy reform and ethics vs law debates.

Explore how ethics and law intersect yet differ, and why an ethical framework guides technology, data, and ai beyond legal minimums. Compare behaviors enforced by law with those not codified.

Explore ethical dilemmas through the trolley problem and its modern twist with self-driving cars, examining how data, AI, and machine learning raise questions about algorithms, laws, and societal values.

Highlight key takeaways on data and AI ethics by urging an ethical mindset to identify risks, safeguard against lapses, and develop guidelines for ethical data stewardship and AI use.

Apply ethical data stewardship by securing consent, protecting privacy and confidentiality, minimizing data collection, restricting access, and safeguarding data with anonymization and strong security practices.

Examine the OkCupid dataset breach, focusing on consent, privacy, and data stewardship. Explore how non-anonymized personal data and ethical data practices shape privacy and trust in the digital age.

Explore the ethics of consent in data stewardship, distinguishing legal versus meaningful consent, and learn how to collect, use, and share data with user oversight and opt-out options.

Protect personal data by applying data security practices, including minimal data collection, encryption, anonymization, strong network controls, and employee training on phishing and access rights.

Protect user privacy and confidentiality by practicing data minimization, ensuring transparency and consent, and applying anonymization to minimize security risks in data handling.

Explore ethical data stewardship of genetic data from the user’s view, examining consent, transparency, and governance when sharing with researchers, law enforcement, marketers, and foreign entities.

Analyze social media data ethically by balancing aggregated metrics with respect for privacy, avoiding personal message scrutiny and private account lookups to maintain user trust and data stewardship.

Explore how data ethics intersects law through HIPAA and GDPR, including privacy, security, informed consent, and the right to be forgotten, plus regulatory penalties and compliance challenges.

Learn the core of ethical data stewardship, focusing on consent, security, privacy, and confidentiality, ensuring informed consent, opt-out options, and safer data handling as you address data and algorithmic bias.

Explore how data bias and algorithmic biases arise from sampling and selection to confirmation bias, and how ethical awareness improves analysis and decision making.

Explore how the eigenface algorithm used linear algebra to identify faces, and how training data biases—skewed toward light-skinned, male faces—led to algorithmic bias in facial recognition.

Learn how training data teach models and test data evaluate accuracy, and how bias in data harms underrepresented groups in tasks like diamond pricing and facial recognition.

Explore how data bias can harm organizations through poor decisions, discrimination, and biased algorithms. See how skewed surveys and feedback loops mislead product strategy and call center support.

Evaluate sampling methods to identify bias risks and ensure representative customer survey data, comparing random sampling, stratified sampling, text message outreach, and in-store sampling.

Examine how biased data, underrepresented groups, and proxy variables contribute to algorithmic bias, and explore how feedback loops reinforce biased patterns in automated decision making.

Examine how Amazon's hiring algorithm used past hires and top performers data, revealing data bias and proxy variable bias that discriminated against women, leading to its scrapping.

Explore how proxy variables, including geographic indicators, can reflect demographics like race or income, and examine how unexamined proxies can introduce bias and discrimination in data analysis.

Explore how proxy variables can encode age in pricing models, with examples like work experience, social media use, and doctor visits, and assess contextual ethics to avoid unfair discrimination.

Examine how domain, transparency, oversight, and scale influence algorithmic harm, especially in high-stakes areas like criminal justice, health care, and hiring, and emphasize fairness and accuracy.

Explore model transparency, contrasting white box and black box models to understand how interpretability reduces unethical decisions; examines regulatory pricing demands and strategies to combat data and algorithmic bias.

Identify bias sources, including sampling and selection bias, audit data and models for representativeness, use stratified sampling when needed, and foster an ethics-driven mindset with diverse perspectives to reduce harm.

Explore how to build a disaster response algorithm that distributes aid fairly, examining data ethics and prioritization choices, economic damage, infrastructure, vulnerable populations, weather, and resource availability.

Explore how recidivism risk algorithms like Compas influence parole decisions, expose racial bias and transparency concerns, and raise ethical questions about using models in criminal justice.

Explore accountability, transparency, and context in algorithmic decisions, and assess when humans should take priority over models.

Examine where to outlaw or restrict algorithmic decisions and which choices stay in human hands. Consider coupon delivery, criminal guilt, and career matching, weighing accuracy against accountability and human oversight.

Explore data bias and algorithmic bias, including how non-representative data and biased training data can cause discrimination, poor decisions, and feedback loops, and why algorithms aren’t always the answer.

Explore the ethical implications of artificial intelligence in daily life, including societal impact, data stewardship, bias, hallucinations, and responsible use in education and content creation.

Explore the ethical and legal issues of training data for generative ai, including intellectual property, consent, and the impact of lawsuits against OpenAI and stable diffusion.

Explore generative AI and its ethical implications, from training data and prompts to outputs like text, images, and audio. Examine IP theft, privacy, bias, and confidentiality concerns shaping AI development.

Examine ethical questions around generative AI reproducing a creator’s style on YouTube, including opt-out, consent, compensation, and potential revenue sharing for model training.

Explore how generative AI generates product reviews and affects consumer trust in online ratings. Assess fraud, fake purchases, and ethical concerns for shoppers and sellers.

Explore how generative ai creates text, images, audio, and code that resemble real content, highlighting hallucination risks, potential fraud, and the need to validate outputs.

Examine ethical questions around AI generated celebrity likenesses and consent, exploring use cases from internal research to replacing actors, and the impact of audience knowledge on data and AI ethics.

Mitigate AI risk by using ethically sourced, consented data and transparent training; rigorously test for bias, monitor outputs, and apply critical thinking and source-based fact-checking.

Recognize the ethical risks and impacts of AI on personal life and society, and emphasize responsible use, validation of outputs by data professionals, and government regulation.