ICO and Alan Turing Institute consult on guidance on “Explaining decisions made with AI”

• United Kingdom

• Technology - Articles
• Industrial engineering
• Industrials
• Technology, Media and Telecoms - Disruptive Technology
• Technology, Media and Telecoms - Technology

05-12-2019

On 2 December 2019, the Information Commissioner’s Office (“ICO”) and the Alan Turing Institute (the UK’s national institute for data science and artificial intelligence) (“Turing”) launched a consultation seeking feedback on their new guidance: “Explaining decisions made with AI”.

This consultation has been developed in response to the UK government AI Sector Deal (published in April 2018) which, amongst other things, tasked the ICO and Turing with developing guidance to assist organisations in explaining to individuals how AI is used in making decisions that affect them.

The guidance is made up of three parts:

Part 1, the basics of explaining AI, focuses on the relevant legal framework that organisations need to comply with and what goes into an explanation. It sets out key principles that should be followed, namely being transparent, being accountable, considering the context in which the organisation is operating and reflecting on the impact of the particular AI system on the individuals affected and wider society.

Part 2, explaining AI in practice, contains a 7 step guide on how to go about creating and providing explanations.

Part 3, what explaining AI means for your organisation, considers at a more strategic level what roles, policies and procedures organisations will require in order to provide meaningful and compliant explanations.

Responses to the consultation are due by 17:00 on Friday 24 January 2020 and links to the consultation and guidance can be found on the ICO’s website: here.

Charlotte Walker-Osborn, International head of artificial intelligence and head of technology sector comments:

“The balance needs to be struck between providing the relevant level of transparency and explanation that lawmakers and consumers might like or need but that technically and from a proprietary viewpoint may be over burdensome. In this respect, it is imperative that both AI providers and customers or beneficiaries of AI engage with this consultation to add their views in at the right stage.”