There’s a lot of buzz in the tech world about ChatGPT. ChatGPT is an AI chatbot that can create human-like answers to any question you ask it.
What the bot said about Eskom
After all the rave posts on LinkedIn, I had to try ChatGPT for myself. It’s spectacularly better than the (awful) chatbots I’ve come across on sites like Fedex.
I asked ChatGPT this question: How do we fix Eskom in South Africa?
It gave me a grammatically-perfect 247-word response, that included statements like this:
Eskom has been plagued by financial and operational mismanagement, which has led to a significant debt burden and poor maintenance of its power plants. Addressing these issues is crucial for the long-term sustainability of the company.
It could have been a speech written for an assignment – or a politician.
War of the bots
Older web developers remember the browser wars. Now the search engine wars have started, and their weapon of choice is chatbots.
Microsoft is adding the ChatGPT model to its search engine, Bing, and its next version of Edge.
Not to be outdone, Google is racing to launch its own chat-engine tool, called Bard.
But the bot was wrong
Perhaps Google raced a little too fast. One of the examples used in the promotion of Bard had factual errors. That may, or may not, have explained why Google’s shares fell 8% after the mistake hit the news.
And that’s one of the problems. AI chatbots, like ChatGPT, sometimes get it wrong. They don’t query a database of proven facts to answer questions. They are trained on huge bodies of text and analyze patterns to determine which words follow the next. (One prominent AI professor described them as "b…s… generators".)
The bot(tom) line for humans
There are and will be many debates about the consequences of AI.
The idiom "Don’t believe everything you read" dates back to at least 1864. It’s more important now than ever before. We know the internet is full of false information, but do we really test what we read? Especially if what we read happens to align with what we want to believe (aka confirmation bias)?
ChatGPT can write a beautiful essay on critical thinking. But relying on tools like this doesn’t teach us how to think critically. And that’s the skill we really need.
Have you tried ChatGPT? What about ChatGPT makes you worried? I’d love to hear your views.