There is no I in AI
Why hiring pros should welcome our new robot overlords.
“Jobs are a bundle of tasks.”
I heard this on a podcast the other night and immediately googled to find out where it came from. A few seconds later I was staring at an econ paper entitled A Task-Based Theory of Occupations with Multidimensional Heterogeneity by Sergio Ocampo.
A 66-page academic paper with an inscrutable title was a little more than I was bargaining for at 9 PM.1
So, without enough gumption to read the paper, I asked ChatGPT to summarize it. Then, at 9:01PM, I realized I was also without enough gumption to read an academic summary. So I asked GPT to whip up something more my speed:
I was torn whether to be more troubled about the child labor analogy or robots coming for those poor children's jobs.2
...
Change is trauma.
When we think about AI, predictions about its future, and how it might affect our current jobs, it can feel pretty darn scary. Of course, GPT dystopian summaries of child workers being defenestrated don't help.
The good news? I don't think GPT provided a terribly good metaphor! Because jobs are a bundle of tasks. And very few people in today's economy are so one-dimensional.
I'll let Arvind Narayanan of AI Snake Oil fame (via the 20VC podcast) speak more eloquently about the fear of robots taking jobs (emphasis mine):
When ATMs became a thing, you know, it would have been reasonable to assume that bank tellers were just going to go away. But in fact, the number of tellers increased. And the reason for that is that it became much cheaper for banks to open regional branches. And once they did open those regional branches, they didn't need humans for some of the things that you couldn't do with an ATM. You know, the more abstract way of saying that is as economists would put it, jobs are bundles of tasks and AI automates tasks, not jobs. So if there are, you know, 20 different tasks that comprise a job, the odds that AI is going to be able to automate all 20 of them are pretty low.
"AI automates tasks, not jobs."
I really like this concept. And I think it helps us imagine a world of job augmentation across the economy, not job replacement. Just like the internet changed everything and the iPhone changed everything, in a few years we'll only have gauzy memories about how we muddled through without AI. But we’ll continue to muddle through.
To unravel the jobs vs. tasks framework in a simplistic way, let's look at what I've done in this particular missive. I had a job—write a Piggyback article.
The job consisted of multiple tasks:
Noodle on the state of the world re: hiring and recruiting
Jot down notes into my to-be-or-not-to-be-article-tickler
Pick a topic
Research and summarize content
Gather some block quotes
Write the article
Publish the article
AI's strength is decidedly not walking in the woods whilst weaving together narratives from the disparate components of everyday serendipity. Just like it isn't very good at writing compelling beach reads. As a probabilistic language model, the arc of its nature bends towards generic. This is good news for humans, because we can often do this non-generic stuff pretty well.
However! AI is pretty frickin' great at doing mundane, time-consuming tasks quickly. And, in more good news for humans, not many people love performing mundane, time-consuming tasks! Like how AI can ingest a 66-page academic paper and spit out a summary in seconds. Or how AI can ingest a podcast MP3 and, in minutes, create a transcript I can copy/paste into this article.3
As I think about AI recruiting in this brave new world, I ponder the value a human can add versus the mundane tasks many of us grind through— just because there is no one else to do it.
I’m willing to take this bet: mundane tasks can and will be automated by fine-tuned bots over time, like scouring through CVs and screeners and interviews in the blink of an eye and creating magical summaries and analysis—like the stuff that Truffle does with one-way interviews.
But then we can ponder the tasks that are not so easily handed to our soulless machine-prediction-friends: Living the intricacies of company culture. Understanding the eccentricities specific needs of hiring managers. Hand-holding a candidate through the stressful applicant-to-candidate-to-new-hire process.
Scott Belsky, EVP at Adobe, gave a talk last year about how AI will transform the creative professional industry. His tidy summary: AI both lowers the bar AND raises the ceiling. For example, on bar-lowering, AI enables an art-impaired person like me to create an image from a few lines of text—in seconds.
However, for the pros? It offers superpowers to create far more in far less time. And for a creative person, this means more cycles to explore and come up with truly magical inventions.
I'll let Scott provide an anecdote about how removal of the mundane can make a huge impact (emphasis mine):
So we actually just announced the other day a new capability in Adobe Illustrator…called generative recolor…If you're an illustrator…a lot of your time is spent trying variations of colors that go well together…You come up with a color palette you want to try pretty quickly, but then the process of actually trying them is excruciatingly long. So here, you can actually put any prompt in, select the entire illustration with one click, and then instantly change all of the color palettes across the entire document within seconds. And when we show this to our customers, some of them actually cry. Like, they literally tear because they're like, oh, my goodness. Like, the amount of my life I would spend just doing various versions of this for my client—or for my internal client or for the poster I'm making or whatever—(is) suddenly eliminated. And then they start to say, look at all the things I can now do with that time. Instead of just exploring color palettes…I can use that time to explore all of the different variations of this illustration. I can really be far more creative in the paths that I'm going to pursue to solve this problem I was given.
For the recruiting and HR pros out there: Close your eyes and imagine a world where you had the 10-20 best candidates handed to you on a silver platter from your inbound funnel. All the noise you experience today does not exist. You are now free to engage with those candidates and sell them on your company and give them the white glove treatment you've always wanted to give them—if you only had the time. People will celebrate you for your candidate experience. As a bonus, you also have time to better serve your hiring managers to help them become better hirers4. You have the time to truly be the Custodian of Hiring, which, for my money, might be the most important thing a company can get right. Your ceiling has been raised.
Now let's look at the not-yet-pros. Consider a new-ish hiring manager at a small company who doesn't have a team to guide them. What if they, too, could close their eyes and start with the 10-20 best candidates? You can see how they’d avoid all the rookie mistakes—heavily leaning on fancy company names on a resume or being blinded by impressive-sounding CV bullets instead of assessing for culture or behavioral attributes. AI won’t replace the pros (but, to be clear, the pros never existed for these hiring managers in the first place!)—but lowering the bar offers a huge opportunity for higher productivity and quality.
Augmentation. Not replacement.
There is no I in TEAM and there is no I in AI. In the not-too-distant-future, it may even become your best teammate ever. Embrace it.
To be fair, it’d be more than I’d bargain for at any time.
I’ve done the “5-year old” trick a couple other times over the past week; once to explain some tax guidance on qualified dividends and another about some intricacies of product marketing—each time it has given me “toy” analogies. When it came to dividends, it was “like a toy that gives out candy every month.”
For transcription, I’ve been using Open AI’s Whisper Transcription app for Mac; it works great.
This word does not roll well off the tongue.