ChatGPT has been the talk of the world for the past few months. Even though AI has been in the works for quite a while now, ChatGPT is the beginning of a mass market set of AI tools available for general use, and widespread use has started to make clear the opportunities and risks that come with it.
While teachers are struggling to grade essays that may or may not have been written with ChatGPT, companies are incorporating AI into their everyday structures, leaving people fearful for their jobs. A friend who is a lawyer for an IT company told me of recent commentary from her engineering colleagues on the lawyers being unnecessary and replaceable with AI.
Software engineers think everyone else is replaceable but them. Lawyers largely think everyone else should be replaced but them. Massage therapists feel secure about being irreplaceable – for now. Doctors and nurses seem in permanent shortage in the wake of the pandemic. Coaches have never been in as much demand as they are right now – how can this continue? What professions are secure? Which ones are not? As AI advances, what can and can’t it replace?
All Hat and No Cigar
Never before has a tool this fast and powerful been made available to such a broad audience. The internet abounds with the polarities of those arguing for and against generative AI.
On first use, it’s hard not to be wowed by simply the ability of Generative AI to take in and respond to specific directions – in that sense the initial impression is beguiling. But just like every new thing, the initial high wears off as you use it more. As you dig into the answers it provides, you realize that it lacks consistency.
Understanding why it picked the information it picked to present has rapidly become beyond the comprehension of the engineers who built these generative AI tools themselves – the inner workings of the large language model become more obscure and opaque with each successive iteration.
For a little while, Generative AI tools made me stop using Google, which seemed so 2022. I switched to using ChatGPT and Bard for a time. Soon thereafter though, I found myself switching back to Google – the black box that spat out cohesive sounding sound bites in answer to my questions seemed less and less reliable. I became fearful of not knowing what I didn’t know. If I could not understand where the AI was gathering the information it was presenting to me from, how could I know if it was valid?
For simple tasks – like writing code, creating simple legal agreements – using an AI tool is a no brainer. No one’s complaining about the simple tasks becoming cheap or free as a result of AI. For some time now, one of the side effects of demand based pricing has been to raise the prices paid to highly sought after professions until the market rebalances with a greater supply of those professionals. This has meant that software engineers or lawyers or doctors or massage therapists that deliver a distinctive, higher quality product that’s hard to replicate, command a premium. It’s hard to see AI substituting those people. It’s easy to see AI making it easier and more accessible to more people to assess when a product or service is actually of distinctly higher quality: Journeymen 0, Experts 1.
Can ChatGPT Be Your Coach?
To help answer this, I asked ChatGPT a simple coaching question: How can I improve my overall work performance?
Here’s the answer I got:
“To enhance work performance, set clear goals, prioritize tasks using methods like the Eisenhower Matrix, manage time effectively with routines and techniques such as the Pomodoro Technique, eliminate distractions, continuously develop skills, delegate and collaborate, seek feedback, stay organized with tools, maintain work-life balance and self-care, foster a growth mindset, celebrate achievements, network, and stay updated in your field. Consistency and a balanced approach are key to gradual improvement.”
The answer is reasonable: it sounds comprehensive, the matrices and techniques it cites sound vaguely technical and all the recommendations encompass the elements necessary. But to those of us who know anything about coaching, it is fairly obvious. The difference between this and the answers a Google search generates is simply that this is an edited compendium of the tidbits that would usually reveal only by looking at the top few Google search results. But is it enough? Is it actually going to illustrate how to improve their work performance to someone and enable them to see tangible improvement?
What Makes AI Different from a Real Coach?
For the vast majority of individuals that seek and use coaching, they already intuitively or explicitly know what to do to improve their overall work performance. The challenge lies in actually being able to do it and the reason they find coaching so deeply powerful is that a coach provides tactical advice on how to do what they know they need to do. Coaches do this by helping each individual build their own specific roadmap of tactical steps to take, hold them accountable to execute them, and help them decomp, improve, and reiterate these tactical steps until they become ingrained behaviors that help an individual reach their goal.
This human aspect is and will always be missing from ChatGPT. AI can provide information, suggestions, and guidance based on patterns in data, but it lacks the emotional intelligence, personal connection, and deep understanding of individual circumstances that coaches provide. Emotional intelligence is difficult to model precisely because it is imprecise – we haven’t yet built a large language model that can be consistently and precisely imprecise.
For example, for me personally, a coach would not have recommended the Pomodoro Technique since I find 25 minute intervals of work too short to get things done. I prefer working for a longer period, and then taking a longer break. No matter how many follow up questions we ask ChatGPT, it can’t tailor the answers to your circumstances or preferences. If we could use a one size fits all model for everyone, we wouldn’t see the differences we see in individual strengths and weaknesses and in the different ways different individuals internalize the same learnings to yield markedly different results.
Here’s Why Coaching is Effective.
Vetted coaches provide coachees with much more than just information. They’re a personal advisor, mentor and cheerleader, helping individuals navigate complex aspects of their lives. Of course they give you the information to work through your problem areas, but perhaps much more importantly, they meet you where you are and work with you to internalize and implement the information so it becomes learning. They do this by understanding your personality, values, preferences, and personal history into account. That’s the singularity in coaching that an AI model has yet to replicate.
Don’t get me wrong, AI will come to play an important role in coaching by helping standardize techniques, helping in the skills enhancement of coaches. But the relationship between coach and coachee relies heavily on empathy, intuition, and a deep understanding of human behavior and emotions. AI can’t do that. The integration of AI tools may change the way coaching gets done, coaches remain irreplaceable for now.
Who would you prefer to be coached by – a sentient human being with their life experience and wisdom or a super smart algorithm who’s providing everyone with the same answers irrespective of who they are? At ideamix, our clients who work with our coaches vote with their feet every day.
