ChatGPT and the future of work
ChatGPT and the future of work
By now, most people have heard of ChatGPT, and many have had a taste of its capabilities. This article explores how this new technology will affect work in the future.
What is ChatGPT?
If you’re still wondering, ChatGPT is a computer system/service that was launched on November 30, 2022. It takes some text prompt and generates eerily sensible “answers”. In fact, it does it much better than any other system we’ve seen before. Together with its little brother Dall-E 2, introduced two months earlier and which can create surprisingly good images on demand, it represents a tipping point for what is called generative artificial intelligence.
- Check this link for a more detailed explanation of generative artificial intelligence, how it compares to other types of AI systems, and how to try it yourself.
What’s different about ChatGPT?
What distinguishes the latest developments in generative artificial intelligence is the quality and relevance of the content produced. Indeed, for the first time in history, it’s possible to see their output as genuinely useful and usable. It’s still far from perfect, but the same could be said of what you get from an average freelance consultant or graphic artist. What’s different today is that using these systems is becoming a viable alternative to employing real people for creative work and research.
Is ChatGPT really that good?
Let’s ask it the question, and you’ll be the judge…
Prompt: Is ChatGPT good enough to replace a copywriter?
ChatGPT’s answer:
- As an AI language model, I am capable of generating text on a wide range of topics and in different styles. However, whether or not I can replace a copywriter depends on the specific task and context.
- In general, I can assist with certain types of copywriting tasks such as generating product descriptions, social media posts, and other types of content that require a straightforward communication of information. However, when it comes to creating complex and nuanced messaging that requires human creativity, empathy, and a deep understanding of the target audience, human copywriters may still have an advantage.
Prompt: Is ChatGPT good enough to replace a research assistant?
ChatGPT’s answer:
- As an AI language model, I can provide assistance in conducting research and generating information on a wide range of topics. However, whether or not I can replace a research assistant depends on the specific task and context.
- In general, I can assist with certain types of research tasks such as gathering information, summarizing research articles, and identifying key points in a text. I can also assist in generating reports, analyzing data, and making predictions based on the available data.
- However, there are certain tasks where a human research assistant may still have an advantage. For example, a human research assistant may be better at identifying and locating sources that are not easily accessible online, conducting interviews with human subjects, or providing context and interpretation of complex information.
The structure of both answers is a bit repetitive, but you can see how each is perfectly adapted to the subtle context difference stemming from its respective prompt.
Ultimately, the decision to use an AI language model like ChatGPT versus a human copywriter or research assistant depends on the specific project and the goals of the content/research. In some cases, it may be appropriate to use a combination of both.
So yes, I’d say it’s already “good enough”.
OK, but what does it mean for my job today?
For now, probably not much. I suspect that, already, some of the recent layoffs at big tech companies are based on the assumption that fewer people will be needed for a host of “knowledge work”, but I couldn’t find any hard evidence of this … yet.
- Have you or anyone you know “lost its job to AI”? If so, drop me a line!
Back to you and your job, though. I like that quote popularized by Bill Gates:
People tend to overestimate what can be done in one year and to underestimate what can be done in five or ten years.
In other words, the changes might be gradual, but they will be significant.
Like Martin Michaud said in a recent post:
- AI will disrupt <knowledge workers’> practice, and they will have to learn to become more “strategist” and less “doers”. <…> Human brains have to reprogram themselves, learn to use artificial intelligence tools and develop new strategies.
Another reference I like to use is that of Steve Jobs, when he refers to computers as “bicycles for the brains”. In keeping with his analogy, artificial intelligence will be the battery that powers your brain bike. If you don’t want to learn and start to use them, you’ll be left in the dust by your competitors!
How will ChatGPT affect my job?
Kevin Scott, chief technology officer at Microsoft, recently gave a pretty rosy outlook on how AI will change the way we work.
- It will unleash our creativity.
- It will make coding much more accessible.
- It will become our co-pilot.
- It will unlock faster iterations.
- It will make work more enjoyable.
I’ll give him a “mostly yes” on creativity. It will unlock creative outlets for people put off by the tedious aspects of, well, creating. But creating truly novel and significant content will still require quite a bit of work and talent.
I have more reservations about coding. I’ve seen too many attempts to “simplify” and “automate” coding (some of which I’m guilty of myself). I’ve seen first-hand how unexpressed or ill-understood requirements can create disproportionate problems down the road. I’ve hit too many times the limits of low-code/no-code systems. Maybe it’s coming from my own professional bias, but I find it hard to believe that even a quasi-magical artificial intelligence can make all the challenges of building modern software go away.
Kevin’s idea of having a co-pilot is, in my opinion, the strongest in his list. It’s akin to the electrical-battery-for-my-brain-bike analogy I mentioned above. I think this is also what can make his next point, faster iterations, a reality.
Will it make work more enjoyable for everyone? I wouldn’t go that far. As with creativity, pleasure is highly personal and contextual. Some will be stimulated and motivated by the prospects opened up by artificial intelligence, but others might never experience that “bliss”.
Now’s the time to get ready!
ChatGPT is real. More broadly, generative artificial intelligence will continue to evolve and improve over time. The changes it’ll bring will happen whether you embrace them or not. You’d better get comfortable with these systems and figure out how you can use them to create more value – for yourself, for your clients or employers, and for society.
Good chatting!
P.S.: Guess what? I’ve used ChatGPT to write or inspire sections of this article, and not just the quotes I directly attributed to it 😊. I used Dall-E 2 for the title image and took advantage of the Office grammar tool and Google Translate. Cheers to a more productive life!