AI and algorithmic incidents and controversies

Artificial intelligence and algorithms can help tackle climate change, strengthen cybersecurity, and improve customer service, amongst all manner of applications.

On the other hand, limitations in fairness, safety and security together with a perceived lack of transparency and accountability means these technologies can damage the rights and interests of citizens, consumers and others and, as such, pose significant potential risks to the organisations that design, develop and deploy them.

As the adoption of AI and AI-related technologies become more mainstream, awareness diversifies and grows, public opinion consolidates and legislation hardens, the risks are likely to become more reputational in nature.

AI and algorithmic incident and controversy repository

Surprisingly little clear, objective, structured information and data exists in the public domain on the limitations, consequences and risks of artificial intelligence, algorithms and automation.

Accordingly, I have developed an independent, open repository of incidents and controversies driven by and relating to ‘narrow’ and ‘general’ AI and AI-related technologies across the world since 2012.

Detailing 450+ negative events, the respository is regularly updated. However, it does not claim to be comprehensive. It does not cover super-intelligence and related meta controversies.

Who uses the repository

The repository is used by researchers, academics, NGOs and policy makers for reference and research.

It is also used by business managers, risk managers, lawyers, reputation managers and others looking to deepen their understanding of AI risks, and to apply their findings.

How to use the repository

The repository is highly flexible and can be used for multiple purposes across a wide variety of settings.

These may range from conducting qualitative or quantitative research; predicting future trends; developing case studies; training and education; or developing and applying methodologies, frameworks and other products and services.

Real-world examples of its use include:

  • to develop and populate an AI incident database
  • to develop a public map of helpful and harmful AI across the world
  • to devise an AI legal response framework
  • to identify and assess the personal risks of AI
  • to analyse the reputational impact and risks of AI, algorithms and automation.

Terms and contributions

The repository is an open resource which anyone is free to use, copy and adapt under the terms of its license.

Accurate, fair and supportable contributions in English are welcome.

%d bloggers like this: