This plain language guide offers a framework for privacy professionals to use when assessing algorithmic systems for privacy risks and harms.
What is an Algorithmic Impact Assessment, and when are they needed? What does a privacy professional need to know about how AI works? What are the different types of bias which should be controlled for? What does a ‘good’ system look like?
For an assessment of an algorithmic system to be robust, it should encompass:
- Legal compliance – ensure the algorithmic system is lawful, with particular focus on privacy, anti-discrimination, and consumer protection laws
- Social impacts – consider the social, political, and economic context for a deeper appreciation of potential privacy-related harms, and
- Technical considerations – integrate testing for accuracy, performance, fairness and bias.
This eBook provides a useful introduction in any jurisdiction. It offers clear guidance on:
- foundational concepts and definitions about algorithmic systems, automated decision-making and artificial intelligence
- the various types of privacy-related harm which can arise from algorithmic systems
- an exploration of how fairness, ethics, accountability and transparency (‘FEAT’) can be built into algorithmic systems
- the Four D’s Framework for assessing privacy risk in algorithmic systems, across design, data, development, and deployment
- a comprehensive list of 66 features of trustworthy systems, which can be used by privacy professionals seeking to assess algorithmic systems, and
- where to place Algorithmic Impact Assessments in the context of other types of risk assessments like PIAs.
This eBook is included with other resources in a number of our value-packed Compliance Kits.
Or click below to purchase this eBook on its own: Algorithms, AI, and Automated Decisions – A guide for privacy professionals.
Questions? See our eBook FAQs.
