I am not shy about the fact I am not a fan of AI (yes, I personally write everything I produce).
AI can do what you tell it to do, what you want it to do, and what you expect it to do, and it can still fail. And that’s before it falls prey to a cyberattack: it works perfectly as it was designed to, but thanks to tinkering or malware, the algorithm produces something intentionally terrible for your business instead.
Business continuity isn’t just for saving your business from disaster, it is also about boosting your business and setting it up for success, both in your vision/mission and reputation. As this is coming out as mental health awareness month just closed, I want to focus on one case study: the VA uses AI to identify veterans at risk for suicide. They think it’s a “game changer.” It’s an abomination and PR disaster for the following reasons:
The VA set up their algorithm to target assistance to patients with the highest statistical risk for suicide. It looks at 61 variables. Preference is given to “divorced” and “windowed” males. No female veteran groups, not even female veterans who experienced military sexual trauma or domestic abuse, are considered.
Is it just math? Or bad programming? Algorithmic bias is a real thing. While the suicide rate for male veterans is higher than female veterans overall, the rate of female veteran suicide is rising faster. The female suicide rate increased 24% between 2020 and 2021. Also, female veterans now make up 30% of patients in the VA system.
Both presidents Trump and Biden issued orders to promote transparency and accountability in AI products. President Biden further issued orders to require federal agencies using AI to make sure it does not disadvantage those already denied equal access and opportunity.
Reputation is an asset that needs to be protected in a business continuity plan and a lens through which impacts are analyzed. Sometimes the math doesn’t math. Be careful of how you can be made to look when you are investing in the shiny new technology toy because it could wind up coming back to bite you.
Also, businesses are spending gobs of money on AI that writes terrible/mediocre emails instead of investing in hiring educated workers to write them instead. Just another example of how it does what you want it to do, but you still fail. At some point you can let go of the kung-fu grip you have on throwing good money after bad.
Comments