Primary supervisor
Waqar HussainCo-supervisors
Disruptive technologies such as artificial Intelligence (AI) systems can have unintended negative social and business consequences if not implemented with care. Specifically, faulty or biased AI applications may harm individuals, risk compliance and governance breaches, and damage to the corporate brand. An example for the potential harm inflicted on people is the case of Robert Williams who was arrested because of a biased insufficiently trained facial recognition system in the US in 2020 (See the New York Times link below).
Student cohort
Aim/outline
AI literature establishes that in order to address ethical issues of AI, software developers need systematic processes and mechanisms to improve the way AI systems are developed. Given the popularity and effectiveness of checklists in improving the state of practice and safety measures in healthcare, manufacturing or aviation, checklists have gained popularity as a means to improve the risk prone process of AI system development too.
The objective of this project is to
- critically analyse existing ethical AI checklists and questionnaire based guidelines
- identify their strengths and weaknesses,
- design an improved (digital) checklist to manage AI training data that improves on the identified shortcomings and
- evaluate the effectiveness of the proposed checklist through developer interviews or focus groups.
The focus of this project will be on ethical issues of data collection and processing activities that are required for training machine learning algorithms.
URLs/references
- New York Times, Wrongfully accused by an algorithm / https://www.nytimes.com/2020/06/24/technology/facial-recognition-arrest.html
- Floridi, L., Cowls, J., 2019. A Unified Framework of Five Principles for AI in Society. Harv. Data Sci. Rev. https://doi.org/10.1162/99608f92.8cd550d1
- Hagendorff, T., 2020. The Ethics of AI Ethics: An Evaluation of Guidelines. Minds Mach. 30, 99–120. https://doi.org/10.1007/s11023-020-09517-8
- Madaio, M.A., Stark, L., Wortman Vaughan, J., Wallach, H., 2020. Co-designing checklists to understand organizational challenges and opportunities around fairness in ai, in: Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. pp. 1–14.
- Mittelstadt, B., 2019. Principles alone cannot guarantee ethical AI. Nat. Mach. Intell. 1, 501–507. https://doi.org/10.1038/s42256-019-0114-4
Required knowledge
Solid Programming skills in ML
Able to work with, critically review and analyse machine learning and ethical AI literature
Basic Knowledge of Qualitative Analysis