Can AI Produce High Value Work Requiring Serious Expertise?

Not At The Scarlett Johansson Level

05.22.2024 I Jeanne Meyer

By now, most of us have been both creeped out and impressed with generative AI’s power to do more. Goldman Sachs predicted last year that companies would use AI to eliminate a quarter of all current work tasks in the United States and Europe. Despite this prediction, conventional wisdom is that AI is not necessarily coming for white collar jobs; AI is merely a tool to handle the mundane, freeing up talent to pursue higher value work. Right?

Sure, AI can help to knock out blog copy or a legal brief, write code for software updates, generate images or develop targeted sales lists. AI tools beyond ChatGPT can be used to reduce time for product development or go-to-market planning, as my colleague from WE ARE THE BOARD Lilly Liu Minkove shares on this TikTok. But over-reliance on generative AI to take on tasks beyond the mundane without proper, expert oversight could lead some businesses to tarnish their brands, complicate customer relationships, reduce product quality or diminish the professional development of their employees.

How to define high value work? High value is where true expertise and special sauce are applied in a transformative way. High value is not generating a draft document that a junior high school student could create. It's not using Midjourney to create an ad campaign that has a wildly off-brand aesthetic, just to save some time and graphic design fees.

Scarlett Johansson

High value is what I’ll call Scarlett Johansson level work: special, expert, trained, in-demand, magical and revenue-generating. Here is a very meta example: There’s an AI generated voice, and then there’s award-winning actor Scarlett Johansson, who Open AI’s CEO Sam Altman repeatedly courted to use her to voice its latest AI assistant technology, Sky. The actor says Altman’s pitch was that her voice would “help consumers to feel comfortable with the seismic shift concerning humans and AI.” ScarJo said “no” for months.  And 48 hours ahead of OpenAI’s unveil of Sky, she said “no” again. Underscoring just how singular Johansson’s voice is, Altman teased out the message “her” on social media; a nod to Johansson’s character in the 2013 movie. But after hearing from Johansson’s lawyers and widespread news coverage, OpenAI has since pulled the Sky voice from Chat GPT. When they couldn't get the real thing, OpenAI tried to replicate it, and failed, creatively and very likely, legally.

Scarlett Johansson level work is what truly seasoned human brains do:  irreplaceable stuff like having the EQ to mentor talent, calm down an unhappy customer, or to pitch and win new business. It’s also that studied eye, that kind of perspective informed with years of training in creative fields, that can reject an AI-generated image that doesn’t achieve the right taste level. It’s that judgement and professional scar tissue that courageously insists that a team go back to the drawing board (and fight for extra budget) because it’s not up to gold standard. Or it’s that data-informed hunch that leads a pro with decades of experience to make a bold, business-building decision.

Another example of how generative AI can’t replace a seasoned human: a client needed to quickly announce a product enhancement ahead of a trade show. It was an extremely busy period, so the client had a junior team member use ChatGPT 4 to create a press release, generate images and a target media list to announce the news. They did not delegate any of the project, including oversight, to their agency or outside advisors. After the release went out, it was sent back by several journalists. One said the announcement was obviously “GPT’d.” Another said the product shot looked “really weird.”  And still other writers fired back asking to be taken off the client’s list because they did not cover anything remotely near that product category. It created a real fire drill and only then did the client ask me to step in to develop a potential crisis response in case their obvious use of AI would become a story (it didn't). Also, the entire effect earned zero media coverage.

Fortunately, there’s a workforce trend that offers a counterbalance to the sort of bland output that often results from generative AI tools:  a growing number of senior level executives are hanging up a shingle and offering their expertise to more than one company at a time. And in response, companies are also leaning into this movement, recognizing it as way to access top talent at a fraction of the cost. Studies show hiring a COO on a fractional basis for just 20 hours a week can save $56,000 annually compared to a full time salary. Deloitte estimates that the average cost savings from employing a fractional sales leader is 36%.

When it comes to going fractional, matching high value projects with high value talent via the right network can make this process exponentially more efficient. My participation in organizations like WE ARE THE BOARD, a collective of fractional execs from different professional functions, has paid significant dividends for me by bringing me opportunities that I would not know about otherwise. And TheBoard’s clients have been able to, as just one example, turbocharge their business development by activating members to provide warm introductions.

And both examples are high value – many of them ScarJo level – offerings that require a human touch.