An OpenAI bot called ChatGPT is causing a stir online after more than one million users signed up in the first few days of its release. Read on to find out what ChatGPT is, how it works, and what it means for the future of tech and employment.
Let’s face it, we’ve all dropped the ball at some point. You open up your Google calendar expecting to find a breezy Friday morning ahead of you, but realise you’ve got an important meeting in less than half an hour.
You’re meant to be presenting an updated company value proposition to the board and you haven’t even started it. In the past, you’d say you need a miracle. These days, all you need is ChatGBT. Freakout averted.
ChatGPT is a prototype AI chatbot that’s capable of understanding human language and generating impressively human-like written text. It’s the latest evolution of the GPT – short for Generative Pre-Trained Transformer – a family of AI that specialises in generating written text.
To give an example, if you log into ChatGPT and input a command of “write a company value proposition for Academy Xi”, ChatGPT will rattle off a written response in about 20 seconds. It will be grammatically flawless and astoundingly detailed.
The possibilities are limitless. You can use ChatGPT to whip up a snappy bio for your LinkedIn page, write a corporate white paper, pen the script for a movie, or even compose a degree-level essay. In short, ChatGPT gives you whatever written content you need, whenever you need it.
Beyond taking the heavy lifting out of writing, ChatGPT has a vast array of applications, including:
In other words, ChatGBT has the potential to simplify all kinds of everyday tasks for a lot of different people.
Trained by AI and machine learning, the ChatGPT system is designed to provide information and answer questions through a conversational interface. ChatGPT’s AI is powered by a form of deep learning called neural networking. Neural networks are a series of algorithms that replicate the neuro functions of the human brain. Each of these neurons:
How data moves between neurons within a network and the calculations performed will depend on what data findings are uncovered along the way. Though a neural network makes decisions about what to do with data all by itself, it first needs to be trained with data inputs. ChatGPT’s AI has been trained with an enormous sample of written text taken from the internet.
The research company who developed ChatGPT, OpenAI, said the new AI was created with a focus on ease-of-use. “The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests,” the company said in a recent statement. While chatbots are nothing new in the tech industry, the sophistication of ChatGPT makes previous iterations seem like child’s play by comparison.
OpenAI is San Francisco-based, led by Sam Altman and receives financial backing from Microsoft, LinkedIn co-founder Reid Hoffman and Khosla Ventures. In what’s been a grim year for the tech sector, with mass layoffs, wrecked stock prices, high-profile cyber breaches and crypto catastrophes, ChatGPT has served as a timely reminder that the industry is still defined by jaw-dropping innovations.
Tech experts and venture capitalists have gushed about ChatGPT on Twitter, some even comparing it to Apple’s release of the first iPhone in 2007. Five days after ChatGPT was launched, Altman tweeted that the research tool had “crossed 1 million users!”.
Back in 2016, tech bluechips like Facebook, Google and Microsoft were talking up their digital assistants as the future of human and AI interaction. They boasted of the potential for chatbots to book Uber rides, purchase plane tickets and answer customer service queries with a human touch.
Six years later, chatbot progress has been slower than expected. The majority of chatbots we interact with on a daily basis are still relatively primitive, often installed on customer help pages and only capable of answering rudimentary questions.
With early ChatGPT adopters demonstrating its ability to carry a conversation through multiple queries, deftly navigating a range of intricacies, the world of so-called natural language processing appears to be entering a new era.
A person wrote this blog. Honestly. But how long before AI is able to write it instead?
In the days since ChatGPT was released, there’s been speculation that professions dependent on content production might be rendered obsolete, including everything from playwrights and copywriters to programmers and journalists.
Academics have generated responses to exam questions they admit would result in full marks if submitted by an undergraduate, and programmers have used ChatGPT to solve coding challenges in obscure programming languages in a matter of seconds.
However, in its current stage ChatGPT is not without its flaws. Its current knowledge base ends in 2021, rendering some queries and searches outside of its expertise.
I asked it to write a blog article about “the future of employment factoring in the impact of ChatGPT” and was served a reply of “ChatGPT does not have the ability to provide information or opinions on current events or trends, or to speculate on the potential implications of new technologies”.
It’s also important to keep in mind that ChatGPT can occasionally give factually incorrect answers and authoritatively present the misinformation as truth, with OpenAI conceding that the AI can sometimes produce “plausible-sounding but incorrect or nonsensical answers”.
While ChatGPT might be able to generate a string of related facts, even presenting them in polished written language, it can’t be counted on to place those facts into dialogue and generate complex arguments. In short, critical-thinking and formulating opinions are not its biggest strengths. Without a doubt, it does have many others.
OpenAI said the platform has “limitations and cannot replace humans”. If you’re in a profession that seems to be threatened by this emerging tech, wipe the sweat from your brow – AI isn’t taking over just yet.
The question remains though, how will ChatGPT be used by companies and people? The tech is versatile enough to be deployed in a wide range of settings to perform all kinds of tasks, so watch this space. One thing’s for sure – whatever ChatGBT is used for is bound to be groundbreaking.