We have an uncensored belief in something we have not entirely understood.
Artificial Intelligence is Assistive Technology
I’m sure most of us regularly see the continued excitement surrounding artificial intelligence in the news, where its potential is regularly bloated and often taken out of context. Your AI-powered algorithm will not dominate and take over the world – that’s certainly not the case. There’s nothing to fear, despite Hollywood continuing to eulogise so-called AI-powered humanoids as nefarious. The latest instalment we saw in the cinema quite recently was M3GAN – yet another cheery movie portraying a robot companion gone astray!
Artificial intelligence today is nothing more than clever programming and smart technology.
Let’s not forget, AI today is nothing more than clever programming and smart technology. It remains an assistive technology to aid both business and industry – and that’s it! However, in a similar vein to the Hollywood doom and gloom theme, our hysteria continues…
Examining ChatGPT’s Knowledge-base
For those of you who have been hiding in fear, Chat Generative Pre-trained Transformer (GPT) or ChatGPT has recently received much attention and excitement – I’m sure most of you have heard of this? Anyway, it is a language model developed by OpenAI and was released to the public last November, as a prototype. In other words, OpenAI will have multiple users testing its software which, in turn, will undoubtedly improve the robustness of the AI-tool. As a former software engineer, I know that when you release software to the public after comprehensive ‘laboratory testing,’ it takes on a whole new life form, and users will test it in a way you didn’t expect. ChatGPT is currently open and free to the public, but OpenAI plan to monetise their service at some point.
So, ChatGPT was developed to converse in a human-like manner with its end-users and was also designed to produce human-digestible content. In short, you enter a query into the AI-powered algorithm, and it parses and processes your request and, more or less, immediately responds with information that it currently has access to in reply to your question. The algorithm is expected to be expanded later this year/early next to accommodate a larger base of ‘knowledge’. This will provide the algorithm with increased know-how, enabling a wider scope and complex set of queries from its users, in turn, offering more comprehensive responses.
It's Just Some Clever Software
Those who have commented and responded to ChatGPT capabilities, have remarked that it is incredibly versatile and adaptive. It’s current acumen is quite broad, to include the ability to not only write source code, but to also debug and correct it – I have to admit, this feature is the only thing that makes me a tad nervous. Likewise, the algorithm has been asked to write music, student assignments, poetry and song lyrics, to name just a few use-case examples. Using similar methods, alongside machine and deep learning, ChatGPT learns and uses ‘supervised training’ techniques to enhance its algorithmic core whilst developing comprehensive datasets. Data remains an integral component in any artificial intelligence algorithm as, without it, AI, wouldn’t exist.
Artificial intelligence is a retronym and instead, should be labelled, as ‘assistive technology’.
Now, with all the comments, articles and videos discussing the ‘pros and cons’ of the software, I have been taken aback by many regarding this ‘end-of-the-world’ rhetoric. In particular, I watched an interview with clinical psychologist Dr Jordan Peterson, where he discussed how powerful the AI tool was and how we should be mindful of its future. Whilst Peterson acknowledged it wasn’t quite smart yet, and likened the algorithm to a ‘Humanities Professor,’ he seemed fearful of its potential. I have to say, I’m incredibly surprised by his reaction, since I thought he may have known better. His scaremongering wasn’t necessary. As an engineer, I like to think I have the advantage of understanding what this tool can achieve and the chaps at OpenAI should be pleased with themselves, as it is some very clever programming. Peterson and others have nothing to be fearful of, as it’s just smart software.
Until next time…
Inevitably, as the algorithm is developed, evolves and becomes widely tested, it will become a very robust and effective tool with knowledge abound. As we already know, a computer can process mathematics, analyse pattern recognition and extract sequences that it was designed to do much quicker than a human. Of course, as the tool builds datasets based on what it has learned, it will become more effective with its results. But ChatGPT is not HAL 9000 (A Space Odyssey) as, like any other AI tool, it’s there to assist, and if a student wishes to ‘cheat’ well, shame on them.
I can describe ChatGPT, as a ‘know-it-all,’ but no-one likes a know-it-all and, after all, for me ‘intelligence’ is the application of knowledge. Oh and, if you’re wondering, I can assure you that I did write this piece, admittedly with the help of my wife, Sarah who was kind enough to proofread my waffle.
So, this is where a “this is definitely my content” Dr G, signs off.
Comments