Twice weekly, I share proven strategies to help data scientists get noticed, promoted, and valued. No theory — just practical steps to transform your technical expertise into business impact and the freedom to call your own shots.
Swearing (Not That Kind)
|
A few years back, I did a piece of data analysis that became the basis of expert testimony in a court case. I was even cross-examined under oath. I understood the gravity of the work right from the start, so I was extremely careful to document and check everything I did. Nevertheless, I was surprised by some of the things that were subpoenaed, in particular scanned copies of the handwritten notes I'd made during meetings with the people who were instructing me. What opposing counsel wanted to know was whether any of them had tried to influence my findings. Fortunately, they hadn't, and the cross-examination went without a hitch. I was reminded of this recently, when friend of the list Paul Rayburn pointed me to a ruling in which a US court ordered an expert witness to disclose the AI prompts used in producing her report. This appears to be the first time this has happened, and it's just one judge's ruling. But it highlights something data scientists should take very seriously. Data scientists go to great lengths to ensure their models aren't statistically biased. Yet we often overlook two of the greatest sources of bias: ourselves and our stakeholders. My handwritten notes recorded what I'd been told and how I'd interpreted it, which is exactly why opposing counsel wanted to see them. They were the only record that showed whether anyone's expectations had biased me. AI prompts now do the same thing, but with a twist. If anyone's expectations find their way into the prompt's wording, they can bias the results, and the prompt leaves a written record that someone else can ask to read. For most data scientists, testifying in court is rare. I've only done it once in my career. Audits and independent reviews, though, are far more common. If you want to influence strategic decisions through your work, your analysis must hold up if someone asks who (or what) shaped it. I never expected my handwritten notes to be read by anyone but me. What would your prompts say about you if an auditor read them? Talk again soon, Dr Genevieve Hayes |
Data Science Impact Algorithm
Twice weekly, I share proven strategies to help data scientists get noticed, promoted, and valued. No theory — just practical steps to transform your technical expertise into business impact and the freedom to call your own shots.
Whenever I go to a new restaurant, for the first few visits I’ll order a different meal every time. Sure, one of them might turn out to be a disaster, but the cost of failure is low and it’s the easiest way to determine which menu items I like the most. This same experimental approach works very well for most of the decisions we encounter every day in our lives and in our work. But what about the big decisions that land on executives’ desks, where the cost of failure is high and there’s no...
We all know the story of Orson Welles' infamous radio broadcast of The War of the Worlds. On Halloween Eve in 1938, Welles dramatised H.G. Wells' novel about an alien invasion so convincingly that listeners genuinely believed the world was being attacked. The result was widespread panic across the US, with people taking to the streets in fear. It's the most famous radio hoax ever, and from the 21st century it can seem pretty hard to believe. And yet we believe it because that's what the...
One of the most frustrating things about working with AI is when you get to the point where it seems like Claude or ChatGPT or your AI tool of choice really understands you and the way you work - and then all of a sudden it doesn’t. It’s like training a new graduate to be useful at their job, only to be told they’ve moved to another role and you have to start again, training someone else. This is annoying enough when dealing with AI chatbots. But, when building AI systems, it can be...