ChatGPT And Reimagining Human Intelligence

By now you have probably heard about ChatGPT (and used it!). Even to those familiar with AI tools, ChatGPT generated a wow moment. Perhaps it is the sheer breadth of possible applications, the accessibility, or the ease of use. In any event, people are scrambling to figure out what it means for them and their businesses. There appear to be three main reactions to date – ignore, ban (with or without detection), or embrace. While completely understandable as short-term reactions, the first two are not intermediate or long-term practical. The technology is too powerful, too easy to use, and too helpful to too many people.

There are many issues and concerns about ChatGPT – such as bias and inaccuracy. This article is not about that. These concerns are legitimate. However, the overarching issue – in my view – is that a trend has started that is not likely to go away. So in the future, we can expect these tools to fix or mitigate as many issues as possible – but retain the functionality and ease of use that has proven so delightful to so many.

A series of tools have sprung up to help us detect whether ChatGPT wrote something. While this is very useful – it is essentially a race – between the AIs that create and the AIs that detect the creator AIs. We see the same race playing out in the deep fakes and misinformation space. However, in that case, it is clear that deep fakes are not a good thing, and positive uses for the same underlying technology (such as medical data augmentation) are easily embraced. ChatGPT (and others sure to follow in its footsteps) raise a more challenging – but not new – question. When an AI easily does what a human is used to doing, takes pride in doing, and has invested time in learning to do, how should society react?

One could argue the examples I show below are not as profound – but we have been here before. What can we learn from the past?

• Before typewriters and laptops, the handwriting was a skill taught in schools. We measured students by the quality of their handwriting and taught skills such as shorthand and cursive. I don’t remember the last time I wrote anything of greater length than a post-it note without using my computer. These days even typing on a computer is less effective than recording voice and editing an AI-generated transcript.
• In the coding world, how we generate programs has changed beyond recognition. Engineers used to write code in assembly or low-level languages. Then came higher-level languages, libraries, APIs, and portable open source. It now takes minutes to generate entire websites that would have formerly taken weeks or months and specialized expertise. I am not including AIs that write code since that is pretty much in the ChatGPT bucket and is the current, not the past.

What can we take away from these? Jobs and the skills needed for them change. But the need for things to be done, and humans to do them, has not changed in these fundamental shifts. They caused massive short-term pain however (for example for anyone who took pains to learn shorthand, or for programmers adept at assembly code).

This brings me to what “embrace” means. Once you accept that ignoring and banning are not practical long-term strategies, we will need to think about what embrace actually means. Ultimately embrace cannot just be about excitement about the technology. It needs to be about how societies and the humans in them ultimately come out ahead because of the technology in question. And embrace means adapt – without adaptation we cannot achieve this objective.

So how to adapt? Depends on who you are. In the first order, I identified the following three groups – for whom adaptation will need to mean very different things

Individuals: Each of us will need to evaluate what skills we have (and rely on) that are already replaced by ChatGPT, or are likely to be replaced once its issues are resolved. Given the pace at which AI technology moves, it is reasonable to assume that known issues will be mitigated at some point. If ChatGPT helps us enhance our value to our environment, leverage it. If it replaces a core value we provide to our environment, it would be wise to look for other core values to develop.

Businesses: Businesses can ask the same question as individuals but of their product(s) value. ChatGPT will launch many new products such as chatbots that will compete with businesses that provide such products today. Again – existing glitches will buy time to adapt, but it is safe to assume the technology will improve. Considering what unique new products and values can be generated by leveraging the technology (and combining it with your deep knowledge of your market segment and customer needs) would be one way to adapt. Businesses would also likely benefit from thinking about how workplace productivity could improve – and how employees can leverage ChatGPT to free up time for other tasks.

Standards-Based Programs. This is the hardest one. While most commercial businesses are measured by what they build, not how they build it (as long as it is legal), standards-based programs such as schools are mandated to teach certain subjects in certain ways. For example, an English teacher is expected to teach her students to write essays, even if now ChatGPT can write such essays. This is also where core skills come into question. Just as we all know that shorthand and assembly language coding are not core skills for today’s students at any level, are the skills we teach now really the core skills of the future? And what will it take to come to an agreement on what the new core skills will be? And what can individual institutions do while the organizations that decide these topics decide what to do? So far, schools and other organizations have opted for the “ban” option – and understandably so since they need time to figure out what to do. Ban does not necessarily mean that embrace cannot start also. It is possible to ban ChatGPT use in some areas while helping students understand what it is and preparing them for a future where it will be part of their life.

This process will be painful, and more painful for some businesses and people than for others. It would be unwise to assume that demand for the skill will return, they never seem to. However, unless we believe that all problems that we need to solve as a society are already being worked on and that all skills we need are already in our collective skillset, there is no reason to believe that human skills have lost value. Human brains are amazingly adaptable but also limited in capacity. If we no longer need a skill, it frees up room for other skills to enter. Every time that a human skill becomes out of date, another takes its place.

To me, really embracing ChatGPT – and all that will follow it – is to redefine what we consider human intelligence. Unfortunately, human intelligence today is measured by a series of tests and skills – these evaluations may no longer apply. We will need to collectively recognize that human intelligence is an evolving concept and should be focused on reaching beyond what any tools of the moment offer, and always looking beyond whatever tools are likely to exist in the near-term and intermediate-term future. What has changed between the examples and the present is the sheer speed at which new technology is arriving and achieving global reach. This will require a much more fluid and agile assessment of needed intelligence – matching future problems to human brains.

Saying this is of course easy. Converting this into real jobs that humans can do to get paid and care for their families is where the problem lies. Hopefully, the conversation turns in that direction sooner rather than later.