18-899-K3   Applied AI for Impact

Location: Africa

Units: 6

Semester Offered: Spring

Course description

Applied Artificial Intelligence for impact (AI4I) offers opportunities to address real-world challenges and generate meaningful economic, environmental and societal benefits. This course takes a practical approach to identifying where AI can create impact and to developing quantitative solutions that improve understanding, prediction and decision-making. Through case studies in healthcare, energy, finance, agriculture, climate resilience and public services, students will investigate how data, models and intelligent systems can support effective interventions. Topics include predictive modelling, classification, time-series forecasting, uncertainty quantification, risk assessment, optimization, generative AI and intelligent decision systems. Students will compare alternative approaches, evaluate limitations and translate model outputs into actionable recommendations. Particular emphasis will be placed on responsible AI, the selection of measurable outcomes, and the distinction between model performance and real-world impact. Practical assignments will involve working with real datasets and communicating evidence-based conclusions to decision-makers.

Learning objectives

The objective of this course is to provide students with practical experience applying AI techniques to real-world challenges and assessing whether the resulting solutions deliver value. The course will emphasize:

  • formulating high-impact AI problems
  • selecting and evaluating appropriate models
  • translating predictions and insights into decisions
  • assessing uncertainty, risk and responsible deployment
  • measuring potential and demonstrated impact relative to credible alternatives

Outcomes

After completing this course, students should be able to:

  • Identify opportunities to apply AI to economic, environmental and societal challenges
  • Select and implement suitable AI approaches for prediction, classification and decision support
  • Evaluate model performance, uncertainty, robustness and appropriate benchmarks
  • Translate predictive outputs into decisions and interventions subject to practical constraints
  • Assess fairness, privacy, safety and governance implications of proposed applications
  • Define meaningful impact metrics and distinguish projected benefits from verified outcomes
  • Communicate technical results, recommendations and limitations to non-specialist stakeholders

Prerequisite

18-785: Data Inference and Applied Machine Learning