ConCReTE Curriculum: Context-Centered Responsible Data Science Training & Exploration
The ConCReTE Curriculum is delivered through the ConCReTE platform (rds-concrete.com)—an interactive, scenario-based learning environment that integrates real-world contexts and datasets into modular Responsible Data Science Opportunities (RDSOs). These RDSOs allow learners to simulate decision-making in realistic AI and data science situations, building practical skills for responsible, human-centered innovation.
RDSOs can be mixed, matched, and tailored to different disciplines and experience levels. As learners progress, they accumulate core digital leadership attributes:
- Agency – Selecting appropriate tools, combining AI and human expertise, and reasoning about design choices.
- Confidence – Experimenting responsibly, evaluating what works (and why), and generating new ideas.
- Accountability – Identifying harms, troubleshooting systems, communicating tradeoffs, and recognizing human–AI entanglement.
| Gain agency by: |
Gain confidence by: |
Gain accountability by: |
- Accurately identifying sets of AI and digital tools that could be used for specific data science tasks that can contribute to accomplishing a business objective in a given role and context;
- Efficiently implementing and organizing different combinations of AI tools and human knowledge to process or plan data processing pipelines; and
- Critically reasoning about why some choices might improve or harm performance and the underlying principles in specific types of applications,
|
- Responsibly and safely trying out and experimenting with new AI innovations and combinations of digital and human contributions that might contribute to relevant business objectives;
- Critically evaluating and effectively justifying when experimental implementations are likely or not to contribute to the business objective and under what conditions; and
- Creatively generating ideas for new experiments from both successful and disappointing experiments.
|
- Identifying harms in the intermediate or eventual output of deployed systems;
- Troubleshooting identified harms by using their knowledge of different choices across AI and human collaborators and digital tools;
- Growing or adapting their teams and processes to identify and deploy solutions to sub-optimal output given the relevant business objects through collaboration with the appropriate AI and human skills;
- Responsibly applying and monitoring insights learned from smaller-scale experiments to deployable systems;
- Communicating the benefits and costs of their solutions relative to other potentially beneficial strategies; and
- Understanding the innate entanglement of AI and human collaborations (bias based on human guidance)
|
