ChatGPT has taken the world by storm and has accumulated millions of users in just a few weeks. Boardrooms all over the world are scurrying to find ways to integrate it into their daily operations, with the hope that it will save them time and money, and probably people too.
What is chatGPT?
ChatGPT is a program that uses advanced computer techniques to understand and respond to human language. It’s been trained on a large amount of text data and can answer questions, generate text and do other language-related tasks. Think of it as a virtual assistant that can understand and respond to natural language. It’s like a computer version of a human who can understand and respond to written or spoken language.
Why is it so popular?
ChatGPT is a popular language model because it has been trained on a large dataset of conversational text, which allows it to generate human-like responses to a wide range of topics. Additionally, its ability to generate text in a specific style or tone makes it useful for a variety of applications such as language translation, question answering, and text completion. Its ability to generate text in a specific style or tone makes it useful for a variety of applications such as language translation, question answering, and text completion.
Where do we go from here?
A million use cases
I already use ChatGPT daily for a variety of tasks. Here are some actual examples of my interactions with it just today:
Asking technical questions
ChatGPT is great are providing summarised answers, unlike Google, which responds with a long list of results, requiring a lot more time to find what you are looking for
Asking it to suggest improvements to existing code
Scanning technical CVs to distill to its essence, then ask it follow up questions
Pitfalls of using this technology
ChatGPT speaks American English. This means its tone is not always clear for non-American cultures and is not useful for non-English speakers (At this point in time). This is likely to be addressed in future versions.
No commercial model
There is currently no commercial or subscription model available. This means that the terms of service could be changed at any time, making real-time integration non-feasible at this time.
OpenAI (the founders of ChatGPT) are working on a commercial model, so this is bound to change.
ChatGPT is in a public beta. This means it is often unavailable due to load exceeding its server capacity. This in turn means we cannot rely on it being available when we want to ask it a question.
My 10 AI principles – a compass for the new world
This technology is here to stay and will forever change how we interact with machines. Working smarter, not harder is the key.
How should we approach this brave new world?
AI and automation will eliminate jobs currently held by humans. We need to tread carefully here.
2. “With great power comes great responsibility”
Using an AI is not a replacement for doing the work. We cannot outsource our jobs this way. Rather, we can automate subsets of our daily routines, and focus on the parts where our critical thinking is required. Getting this balance wrong has the potential of “dumming down” the workforce and losing the key edge that some people contribute to a job.
3. It is not a competitive advantage
This technology is available to everyone, and everyone will be adopting it. It will not give your company an edge. Not adopting it will give your competitors an edge. Consider it a race, with everyone just out of the starting gates.
4. Privacy matters
It remembers everything and sharing personal or business critical information with it may result in data leakage, so take care what is included in your requests.
5. Understand the commercial model
Once the commercial model is in place, costs and rate limits will need to be investigated before we can roll this out on a big scale
6. No more mundane tasks
ChatGPT is great at eliminating or greatly speeding up mundane tasks, like code reviews, scanning CVs or writing text for marketing proposals. It is also apparently great at analyzing data sets, although I have not yet experimented with that.
We need to experiment and learn to take the best advantage of this new technology. Encourage your teams to give it a try.
7. Remember the 80 / 20 principle
AI is not here to replace humans. Yes, we will be able to automate some jobs, like basic customer service agents, etc., but AI will not do all the thinking for us.
For example, it cannot write code for our business, as it does not hold the big picture of what our business does, nor does it know all the systems we integrate with. It can write small pieces of code, but a human with the big-picture view needs to review this code, and stitch it together to make it work in a big application ecosystem.
8. Learn new skillsets
The new skill to have will be to know what to ask. This requires experience and critical thinking., which cannot (yet) be outsourced to a machine.
9. Context is everything
Simply copying and pasting the output from an AI is dangerous and must be discouraged. Context is everything, and ChatGPT does not have enough of it.
10. No one knows everything
This is an evolving landscape, and we are all learning and finding our feet. Be prepared to adjust course as this space matures.