Zoom CEO Eric Yuan has speculated that in five or six year’s time we will all have AI “digital twins,”
To just pause a minute – these are people that are so capable they can attend meetings for us.
To get a feel of what this might be like, check out this video of Reid Hoffman, co-founder of LinkedIn, having a back-and-forth conversation with his own AI digital twin.
While it’s all very spooky, what this video demonstrates, is that we do need to think hard and figure out (fast), if digital twins will play a role in the future of HR.
The term “digital twin” is intriguing but it is simply another term for a computer model.
A facilities planner might create a computer model of a real building (ie a digital twin), so they can experiment with different layouts.
A traffic planner might make a computer model/digital twin of roads in a city to find ways to improve traffic flow.
In the world of humans, you might train a large language model (such as GPT-4) on everything you’ve ever written (which is what Reid Hoffman did), and then others can ask questions of the computer model when you are not around.
The trouble with the digital twin metaphor is that it takes us into a sci-fi mindset far beyond what is possible today.
We’ll do better if we step back and think in terms of specific tasks that could be potentially automated with an AI.
In particular, if we are starting from the digital twin metaphor, we are thinking about tasks that require some knowledge of our style and preferences.
We have seen an example of such a task in our earlier post about the Evie.ai interview scheduling tool. It is not a generic tool, it makes decisions based on your particular processes and preferences.
This isn’t an example of a digital twin, but it is an illustration of how AI can take on a task that needs a fairly sophisticated understanding of your – or rather the organization’s –preferences when it comes to scheduling interviews.
Let’s consider the application Yuan has in mind: an AI application that can replace you in meetings.
Well, there is already a tool that won’t replace you but may still allow you to reduce the number of meetings you attend: automatic transcriptions and summarization.
There will be many cases where you want to know what went on in a meeting, but there is no need for you to sit through the whole thing.
In this scenario, you look through your calendar and tag the meetings where you’ll have little to add and you just want the output. You can ask the organizer to enable an AI to generate a transcription and summary. This will save you a lot of time, almost like allowing an AI to attend some less important meetings on your behalf.
A second type of meeting might need occasional input from you, without a pressing need for you to sit through the whole meeting.
In this case, you could again enable the transcript and summary which you would peek at every 10 minutes or so while you do other work.
If something comes up that you want to weigh in on, then you could send an instant message to the meeting organizer and share your two cents.
Similarly, if those in the meeting felt they needed your input they would know you are on “standby” and are available to immediately answer questions even though you are not sitting through the whole meeting.
Yuan envisions an AI Digital Twin that knows you well enough that it could generate responses for you.
What’s possible right now would be to have an AI agent be aware of the key topics you want to monitor and send you an instant message should they come up in a meeting you didn’t want to sit through.
It’s a long way from Yuan’s vision but it takes us towards his objective of attending fewer meetings and it works with today’s technology.
The idea that a digital twin would make a decision on your behalf – “Yes, the team can have two weeks off in Hawaii” – is still farfetched.
It seems far more reasonable to focus on the tasks we can pretty much do with current technology, rather than think in terms of the far-out capabilities.
Using AI tools to cut back on the time spent in meetings is technologically easy to do.
The hard thing will be deciding which meetings you can skip altogether and just rely on the summary, which ones you need to be on standby, and which ones you need to attend in the old-fashioned way.
The even harder thing is likely that we are all so entrenched in the idea of “let’s just invite everyone even remotely involved in the project to the meeting,” that it’s a difficult habit to break.
This kind of change is also hard because it’s not clear who should be driving it.
Should IT be pushing it as a new technology play?
Should HR be leading this as an organizational effectiveness initiative?
Is it up to departmental or divisional leaders to decide this is something they want to do in their area of responsibility?
Everyone is busy and this kind of innovation doesn’t fall cleanly into anyone’s key responsibilities.
But…I’d encourage HR leaders to take the lead in running some experiments to at least see if it’s possible to use existing AI tools to allow people to attend fewer meetings without unwanted side effects.
If you can figure out how to do this, then that’s a big plus for efficiency and the employee experience.
Yuan may be right that 10 years from now an AI twin can replace us at meetings all together, but when that day arrives, they won’t just be attending our meetings, they’ll be doing our entire jobs.
