If you’ve ever faced an issue with an online order — perhaps a food delivery running late — it is likely that you’ve encountered a “chatbot” as you sought redressal for the issue on an app. In that sense, many of us are familiar with the concept of what a “chatbot” is and what it is supposed to do. But this year, OpenAI’s ChatGPT turned that simple experience into something entirely different. ChatGPT is being seen as a path-breaking example of an AI chatbot and what the technology could achieve when applied at scale.
ChatGPT is a chatbot built on a “large-scale transformer-based language model that is trained on a diverse dataset of text and is capable of generating human-like responses to prompts.” The last bit was written by ChatGPT so excuse the jargon-filled definition. But then, is there a better tool to explain what ChatGPT is capable of, other than the chatbot itself? A conversation with ChatGPT is like talking to a computer, a smart one, which appears to have some semblance of “human-like” intelligence.
As someone who reports on technology, I find ChatGPT to be a fascinating tool. And that’s not because I think it is highly accurate or path-breaking. But because it gives a glimpse into the future of AI and the kind of role this tech will play in our lives. This is not a future that could take place 100 years from now. Rather, it brings up more immediate questions that ChatGPT poses.
So, will AI replace all of our daily writing? I sincerely hope not given what I’ve experienced while playing around with ChatGPT. For one, it is not entirely accurate — something even OpenAI has admitted. It is also evident that some of the essays written by ChatGPT lack the depth that a real human expert might showcase when writing on the same subject. It doesn’t quite have the nuance that a human would often be able to provide. For example, I asked ChatGPT how one should cope with a cancer diagnosis. The responses were kind but generic. The type of responses you would find in any general self-help guide. Contrast that with how some survivors have written their stories and you know that AI has a long way to go. After all, it doesn’t have the same experiences as a human.
And what about code? ChatGPT is writing basic code after all. As several reports have shown, ChatGPT doesn’t quite excel at this yet. But a future where basic code is written using AI doesn’t seem so incredible right now.
Will it write college essays, thus making homework, even exams redundant? I hope no college or school student is using this right now. Thus far, a ChatGPT essay is “readable”, but it is the kind of essay college students submit when they are “faffing”. You know, the one where the arguments are specious, the facts are just about there but it lacks depth.
Then there’s the question of how this AI is built. Can we ensure it doesn’t suffer from the same biases and bigotry many of us possess? Can we ensure it is not racist or sexist or discriminatory in its responses? And what about misinformation? ChatGPT has not proven entirely immune to this. These are all challenges which have not been solved.
Still, ChatGPT is a stark reminder of AI’s potential and threat. Eventually, AI will take over and we don’t quite know in what form. But it is clear that it won’t just be a tool to help us. Will “true AI” become a reality, surpassing humans, achieving “sentience” and becoming capable of feeling? Is there even an understanding of how society will change, evolve or decay as AI becomes more ubiquitous and sophisticated? These are the questions we must strive to ask ourselves now, more vigorously than we have been, before it is too late.
.