Aurora50 Logo

AI risk: Board directors need to assess ‘distance from harm’

Aurora50 Pathway20 AI workshop - group photo
Diana Wilde 20 July 2023
Share
 
Last week we held a fascinating Pathway20 workshop in Abu Dhabi on artificial intelligence (AI) for board directors.
We were advised by our excellent speakers
that everyone should stay curious when it comes to AI, even though many people are still very nervous of it and its risk to jobs.

Data-driven decision-making key to AI for board directors

In polls we ran, half of attendees said they used AI daily already.
Of those:
  1. 88% use auto complete in Google searches
  2. 71% use website chatbots
  3. 65% use Face ID
  4. Almost half use voice assistants like Alexa and Siri
  5. Almost half use Uber
  6. 6% use a robot vacuum cleaner or a Tesla!
Three-quarters said the key skill for board directors when it comes to AI is in data-driven decision-making: in being able to interpret and leverage data insights effectively.
And the most important question a non-executive director could ask, according to three-quarters of our workshop attendees, is: “how is the organisation incorporating AI into its overall strategy and business model?”
Half also thought a NED needed to ask:
  • How does the organisation address bias and fairness concerns in AI algorithms and decision-making processes?”
and
  • “How does the organisation ensure that AI-based decisions are understandable and auditable?”

AI & risk: The distance from harm

The ‘distance from harm’ was a phrase that resonated for me in assessing AI.
When a board looks at AI risk to consider any guard rails and governance required, they should use a Red, Amber, Green (RAG) system that incorporates the distance from harm.
For instance, using an AI chatbot on a site is low-risk, as it only uses pre-programmed answers, while the AI being used in medical laboratories is much higher risk.
Another AI tool I found useful was the landscape analysis that was discussed – input, output, impact – that is to say, review the input pf AI (the quality of data, any bias), the output (how reliable is it?) and the impact.

AI prompts for board directors

I thought that a couple of suggested AI prompts for board directors were very useful: you can ask ChatGPT, Bard etc:
  • How would someone argue against this point of view?
  • What are diverse points of view on [this topic]?
We also noted that attendees were questioning the impact of AI on jobs across the business. A new report from the Organisation for Economic Cooperation and Development (OECD) came out last week, too, and found that more than a quarter (27 percent) of jobs in wealthy countries are in industries where AI could replace many human workers.
A big thank you to Abu Dhabi Investment Office and Hub71 for supporting and hosting us at this workshop.