AI is Not a Threat to the Board; Poor AI Governance is
AI has entered the boardroom, front and center.
For directors, it is not an IT issue. It is a governance, ethics and judgement issue.
Board members can ignore it for a little while, but they do so at their own risk.
And the whole issue around AI is about managing risk. Managing risk in an environment where the rules are still being written. Board members, like it or not, are going to have to drive on the AI road while it is still being paved.
Can AI answer questions? Sure it can but that is not where the risk lies, other than the answers be hallucinations. The risk is when makes decisions or aids in the decision making process.
When you feed data to your favorite AI are you breaching an NDA or exposing trade secrets?
Boards have to ask questions like what is this AI doing with our data? Is there a way to audit it? Is it biased? And, from a legal standpoint, who is liable for its answers. I doubt “the AI made me do it” will stand up in court.
One of the places where AI is absolutely being used massively today is in IT and information security. You literally cannot buy an IT or security product today that isn’t hyping it’s AI features. And that is not the only place we are seeing it. Finance, HR (talk about lawsuits looking for a place to happen), Customer relationship management, manufacturing, sales – they are all AI enabled now.
Here are the key points for board to deal with:
- Data privacy and security
- AI literacy and training
- Loss of human judgement
- Bias, accuracy and explainability
- Ethical use and governance
- Regulatory and legal compliance
- Integration and operational barriers
- Information overload and AI paralysis
We are already seeing problems (also known as lawsuits and potentially class actions) in AI enabled applications. Here are just a couple.
- Amazon’s hiring algorithm penalized females because of biased historical data
- Zillow used AI to price houses it bought to fix and flip. The algorithm screwed up so badly that Zillow was unable to recover and shut down that product line at a massive loss
- IBM’s Watson provided unsafe recommendations for oncology treatments. Hopefully, the human doctors didn’t listen to it.
So what does a board (or, if you are a small company without an active board, the senior leadership team) need to do?
Number one – build AI literacy across the board. And, I don’t mean reading a couple of articles in a journal and declaring the board literate. I mean FLUENCY.
Number two – Integrate AI into strategic planning
Third – embed governance into the workflow – policies, humans in the loop for high risk decisions, regular audits for ethics and bias assessments
Next, manage reputation and risk – AI failure is a reputational risk. Not to mention the whole raft of laws that you need to keep track of.
As AI transforms business, boards need to act – now. AI education, ethical guidance, oversight.
And, if you need assistance, please contact us. Don’t be the next company that winds up with a massive problem because you didn’t act now.
Credit: Directors and Boards
by 