Ask a Data Ethicist: What Are the Ethics of Creating Automated Data?

I recently received two “thanks for attending” emails from two different events I registered for, but was not able to actually attend. Clearly, these are auto-generated emails sent to the entire registration list. These emails got me thinking about this question…

What are the ethics of creating automated data? 

Automated thanks? No thanks.

Why do we send thank-you notes? The obvious answer is to convey our actual gratitude! Genuine thank-you notes are lovely, and when you receive a truly heartfelt thanks, it’s an incredible feeling. 

I was fortunate enough to get a very kind note from a student recently, sharing how she was attending a recent talk where the speaker made reference to something we had covered in class, and how much the course material connected for her in that moment. She wrote to say she was thinking about me and the class. It made my day.

A perfunctory, follow-up “thank you for attending” that is wrong, has the opposite effect. It makes the sender look incompetent and careless. Both of the organizations that sent these notes – a university and a not-for-profit – are clearly not aiming to antagonize anyone. Yet, in trying to be efficient by pre-populating a list into an automated system in an effort to acknowledge and perhaps further engage me, they instead left me with a negative impression. 

A False Narrative

In some way, these systems have created a false narrative about me. Somewhere in the digital landscape, my attendance has been incorrectly recorded as being at these events. This is “sort of” a data quality issue, but not the usual type we might think about. It’s not that data was entered incorrectly or needs to be updated. It was false data created by the system itself in the name of automating a process, and it has no basis in reality

Yet, if this system’s record of my attendance and the corresponding emails it sent me are regarded as some kind of truth and used as further input for some other process – perhaps by agentic AI or another automated system – that false narrative is perpetuated and could generate further inaccuracies. I’m not too worried about these particular two events, but I see the potential for this type of automated data creation to go off the rails in more consequential situations. Given the prevalence of data as evidence and automation bias – our propensity to over-trust machines – if the database “says” you were there, what proof do you have that you were not? 

Lazy Communications

The sad part is that this whole situation was very avoidable. A simple cross-reference of either the email logins for an online event or the checklist at the door for an in-person event is all that was needed to segment the attendees vs non-attendees, and send the right message. A simple “sorry we missed you” would have been appropriate and a nice way to keep in touch. 

Send Me Your Questions!

I would love to hear about your data dilemmas or AI ethics questions and quandaries. You can send me a note at [email protected] or connect with me on LinkedIn. I will keep all inquiries confidential and remove any potentially sensitive information – so please feel free to keep things high level and anonymous as well. 

This column is not legal advice. The information provided is strictly for educational purposes. AI and data regulation is an evolving area and anyone with specific questions should seek advice from a legal professional.