On Friday, Professor Scott Galloway released ProfG.ai, a “digital twin” trained on his writing and capable of answering questions in his voice.
That’s just one of many experiments with different models on how generative AI can impact the future of work.
A few weeks ago, BCG released the results of a fascinating Harvard-led study that researched how more than 750 BCG consultants worldwide use GPT-4.
BCG found that generative AI adoption was a “double-edged sword.” There was high variability depending on how generative AI was used by consultants and type of task. In creative product innovation, GPT-4 boosted performance by an average of 40%. In business problem solving, GPT-4 dropped performance by 23%.
What’s more, consultants don’t appear to have a sense of when GPT-4 is helping or hurting. As BCG put it:
“People seem to mistrust the technology in areas where it can contribute massive value and to trust it too much in areas where the technology isn’t competent.”
BCG also found a “creativity trap” even when consultants used GPT-4 for tasks that showed the greatest performance gains. Despite individual improvement, collective creativity was lower. The homogenous output of GPT-4 reduced “diversity of thought” by 41%.
A bigger potential risk is what the lead Harvard researcher in the study, Fabrizio Dell’Acqua, described as “falling asleep at the wheel.” He said that people using AI tend to “switch off their brains” and “outsource their judgement to AI.” This could lead to atrophying skills: 70% of the BCG consultants in the study worried that using GPT-4 would “stifle their creative abilities over time.”
As one BCG consultant put it:
“Like any technology, people can rely on it too much. GPS helped navigation immensely when it was first released, but today people can’t even drive without a GPS. As people rely on a technology too much, they lose abilities they once had.”
There are two emerging models to adopting generative AI at work: the Centaur and the Cyborg. The Centaur model draws a clear line between human and machine work for different tasks. The Cyborg model integrates human and machine tasks throughout.
It’s hard to believe it’s only been ten-and-a-half months since OpenAI released ChatGPT. It will take time and experimentation to figure out the best ways for organizations to adapt.
One of my favorite quotes from Scott Brinker holds true in this age of AI and the future of work:
“Technology changes exponentially; organizations change logarithmically.”
Here are a few related cartoons I’ve drawn over the years: