By Ian Duckles
It happened! I had heard others talking about it, but it finally happened to me! I was in a Zoom meeting and someone had sent along their AI assistant because they were unable to attend.
The program in question, OtterPilot, is part of a suite of offerings from Otter AI, and is just one of the many new tools being offered as part of AI’s promise to improve our existence by saving us from the mundane drudgery of our work lives so we can focus on the really important stuff. As their marketing notes, “Joins Zoom, MS Teams, and Google Meet to automatically write and share notes.” Other marketing on their website focuses on how this tool can save us from a variety of tedious tasks. In thinking about this product, several things immediately came to mind.
First, I wonder how long it will be before I jump onto a Zoom meeting and there are more AI assistants present than humans. At this stage the product is merely a passive listener, but perhaps in the future, all our meetings can be handled by AI and we never need to go to one again (I know this is about the AI dystopia, but maybe there is some utopia here after all?)!
Second, how does this simplify my life? Now, instead of attending a meeting, I have to read through a transcript/summary of the meeting that I was otherwise too busy to attend. At this point, if I am double-booked for multiple meetings, I can no longer bow out of that meeting, but instead will be expected to send my virtual assistant and still read up on all that happened.
Thus, instead of simplifying my life, it actually seems like I have increased the amount of work I have to do. We like to say there are only so many hours in the day and you can’t be in two places at once, but if my AI assistant can cover for me, I can potentially attend multiple meetings at once throughout the day. Sure, I will need to spend a bunch of time after hours reviewing all the material I missed, but that is a small price to pay for omnipresence.
Another issue that struck me about this is the legality of these “solutions.” I live in California which is a two-party consent state meaning that for any recording of a conversation, both parties need to consent to being recorded. According to conversations I have had with lawyers, a violation of this provision can carry a $10,000 fine.
While the specific details of how the OtterPilot works are not explained, the program obviously needs to record the conversation so that it can translate that into machine-readable information that the software can process. It is certainly possible that the recording is deleted immediately afterwards (though given how these companies work, it is probably just converted into additional training data for the program or sold to other developers to train their AI’s), but still, when I am in a meeting with this program I never consented to be recorded. I did recently ask a lawyer about the legality of AI summaries generally, but they honestly weren’t sure how the law applied in these cases.
More generally, I think I should have some say about whether or not I am being recorded. Further, if my suspicions about how this data gets used to train other AI’s is correct, I’m not sure I want a company to use me as a tool to increase their profits when there is no benefit to me. I have lots of problems with surveillance capitalism, but at least Google gives me something in exchange for all my data. In the case of someone else’s AI assistant recording my contributions to a meeting, I get nothing.
My final concern gets to the core of a major problem I have with AI. The promise of AI is that it will save us from tedium by automating much of the drudgery of modern work. While this sounds nice on paper, I have a better idea: let’s just get rid of all the drudgery! If we have too many meetings that we can’t all attend in person, let’s get rid of some of those meetings. If the meetings are really important but I can’t attend them all, maybe we need to hire more people to do all that important work. If some task is tedious and mind-numbing, rather than automate it, maybe we could ask if that tedious task is actually necessary.
Rather than automating bullshit jobs, perhaps we should just get rid of the bullshit.
Ian Duckles teaches philosophy (until AI replaces him) at San Diego Mesa College and is actively involved in his local union, AFT Local, 1931.