Mark this lesson as complete

Finish all lessons and pass the quiz to earn your official course completion certificate.

Bias and misinformation

Video Module 7: Ethical & Responsible Use of AI

Algorithmic Bias:-Bias in AI isn't usually the result of "malicious" coding; rather, it’s a reflection of the data used to train the model. If the training data contains historical prejudices or lacks diversity, the AI will inevitably replicate those patterns.
- Data Bias: If a hiring AI is trained on resumes from a company that historically hired only men, the AI may learn to penalize resumes containing the word "women’s" (e.g., "women’s chess club").
- Representation Bias: Facial recognition systems often struggle with higher error rates for people with darker skin tones if the training datasets are predominantly composed of lighter-skinned individuals.
- Confirmation Bias: AI recommendation engines can create "filter bubbles," showing users only content that aligns with their existing beliefs, which limits exposure to diverse perspectives.
Misinformation and Hallucinations:- AI models are "probabilistic," not "database-driven." They predict the next most likely word or pixel, which can lead to the generation of false information presented with extreme confidence.
- Hallucinations: An AI might invent a legal case, a historical date, or a scientific study that sounds perfectly plausible but is entirely fictional.
- Deepfakes and Synthetic Media: AI can generate highly realistic images, audio, and video, making it difficult to discern reality from fabrication. This poses a massive threat to public trust and political stability.
- Rapid Scaling: Unlike a human troll, an AI can generate thousands of unique, persuasive articles or social media posts in seconds, making it easy to flood the internet with "fake news."

Strategies for Responsible Use:-To navigate these issues, a framework of Responsible AI is necessary for both developers and users.

Pillar	Description
Transparency	Disclosing when content is AI-generated (e.g., watermarking) and explaining how models make decisions.
Accountability	Establishing who is responsible when an AI system causes harm or provides false information.
Inclusive Design	Using diverse datasets and involving a wide range of stakeholders during the development process.
Human-in-the-Loop	Ensuring critical decisions (medical, legal, financial) are reviewed by humans rather than fully automated.