AI blindspots healthcare

This card set will help you avoid the harmful impact of introducing AI systems in the healthcare context. Responsible AI innovation is always important, but especially so in the healthcare context, given the sensitive nature of data and the vulnerable status of patients.

Every person involved in the AI healthcare chain plays an important role in dealing with these blind spots, and therefore this card set can help doctors, nurses, managers, IT staff, and ethics committee members to proactively identify blindspots and provide strategies to deal with them.

The card set enables you and your team to create more reliable and inclusive AI systems. The tool helps you to:

• identify challenges before they cause harm;
• address ethical issues;
• discuss them within your team and formulate actions;
• communicate correctly with target groups and patients.

The card set consists of 2 types of cards:

• The AI Blindspots cards are divided into two groups, namely ‘general blindspots’ and ‘specific blindspots’. ‘General blindspots’ are blindspots that it is useful to discuss with everyone working within the care circle. The ‘specific blindspots’ do not require everyone to be at the table, as these cards are more interesting for specific profiles (e.g., strong focus on legislation). Each card presents a blindspot and contains:
  • A set of questions that allow you to reflect on possible ‘blindspots’
  • A ‘use case’ to demonstrate the importance of that ‘blindspot’
  • Some tools and tips that can help you detect/reduce the ‘blindspot’
• The Ethical Dilemma cards are examples of situations where an ethical problem arises.

The AI Blindspots cards are divided into two categories, i.e., 'general blindspots' and 'specific blindspots'. 'General blindspots' are relevant and useful to discuss with all healthcare providers. For the 'general blindspots' we identified the following AI Blindspots: 

• Suitability

• Explainability & transparency
• Trust
• Willingness
• Diversity
• (Digital) Inequality
• Trade-off
• Consent
• Datafication of health data
• Accuracy & quality
• Impact on work(floor)

• Undesirable impact
• Ownership

‘Specific blindspots’ are the second group of cards. These cards are more interesting for some specific profiles, and on each card, which profiles should best have a look at the card is mentioned. For the ‘specific blindspots’, we identified the following AI Blindspots:

• Target group: management
  • Broad discussion
  • Education & sensibilisation
  • Large selection
  • Changing purpose

• Target group: management & IT
  • Standardisation
  • Data minimisation
  • Security

• Target group: IT
  • Availability of data
  • Contextual factors

• Target group: doctors, nurses, & management
  • Human vs machine
• Target group: ethical committee & management
  • Data governance & privacy

• Download the card set and get started!
• Also, download the PDF below with 2 templates for using the AI Blindspots healthcare card set. In the first template, you start from an ethical dilemma and work with the AI Blindspots healthcare cards (workshop methods 1 and 2). You can use the second template for the reversed brainstorm with the AI Blindspots healthcare card set (workshop method 4). The PDF includes a completed copy each time as an example.
• Would you prefer a detailed step-by-step plan for organising an in-house workshop on AI Blindspots? Then be sure to download the roadmap.

This card set is inspired by AI Blindspot of Ania Calderon, Dan Taber, Hong Qu, and Jeff Wen, developed during the Berkman Klein Center and MIT Media Lab’s 2019 Assembly program. The AI Blindspots card set is available under a CC BY 4.0 License.

The Knowledge Centre Data & Society adapted the original card set to the Flemish context to support the development of trustworthy AI in Flanders.

In a later phase, the Knowledge Centre Data & Society teamed up with VIVES Hogeschool to create a version of the AI Blindspots card set specifically for the healthcare context. VIVES Hogeschool worked on this AI Blindspots card set within the project ‘AI in hospitals: ethical and juridical aspects’, funded by ESF Flanders.