Card set exposes generative AI's blind spots

Generative AI has rapidly made its way into offices, laboratories and meeting rooms. The technology makes work faster and more efficient. But it is precisely this speed that can also conceal risks. Anyone who uses AI without a critical eye runs the risk of overlooking important ethical questions.

To help organisations with this, the Knowledge Centre Data & Society developed the GenAI Blind Spots card set. The tool aims to encourage teams to reflect on the less visible aspects of generative AI and to engage with the technology more consciously.

Each card in the set highlights a potential blind spot. They each start with a central question, accompanied by a brief explanation, a concrete example and reflection questions. This sparks discussions on topics such as data quality, the ‘shadow use’ of AI tools – where employees use AI outside the organisation’s oversight or policy – or the reliability of AI-generated information.

The cards are organised into four domains: the development of AI applications, their use within organisations, the potential consequences of AI output, and the broader impact on organisations and society. The aim is simple: to help teams identify risks more quickly and respond to them in a considered manner.

In developing the card set, the knowledge centre combined insights from workshops with experts, input from professionals in AI training programmes, and academic literature on the ethical challenges of generative AI.

The GenAI Blind Spots card set can be downloaded free of charge from the website. Those who prefer to work with physical cards can also order the set via the VUB Webshop for €10 each. The card set is available in Dutch and English.