Imagine someone bolting two doors and an engine onto a shopping cart and proclaiming they’ve got themselves a ‘fast car,’ absurd right? In a world where buzzwords are as ubiquitous as that one joke everyone can’t seem to miss, let’s address the latest hype on everyone’s lips: ChatGPT.
Before we start arguing about what is and what is not, let’s discuss what is; and what really is AI. AI is not a thing, it’s a broad field of study made with multiple disciplines. Now hand to heart, the subfields of AI are as varied as the number of passwords I’ve forgotten – a lot! But for the sake of progress, let’s keep it simple, we will summarize it to:
- Machine Learning
- Computer vision
- Natural Language Processing
- Fuzzy Logic
- Expert Systems
The list can grow bigger and wider depending on your level of interest.
Now our chatty grey and trusted friend uses a combination of Natural Language Processing, Expert Systems, and lots of Machine Learning.
What’s Under the Hood
Bear with me as we need to get technical a bit, ChatGPT is a Large Language Model (LLM) that uses a type of Machine Learning called adversarial learning (there are more types of Machine Learning than there are ways our finance minister can conjure up weird financial policies). To my wide-eyed wanderers, here is a taste:
- Supervised Learning: Teaching the computer with examples, like showing it pictures of cats and dogs to help it recognize them later.
- Unsupervised Learning: Letting the computer find its own patterns, like sorting a box of assorted Lego bricks into different piles without labels.
- Semi-Supervised Learning: A mix of teacher and self-study, where the computer learns from both labelled and unlabelled examples.
- Reinforcement Learning: Teaching the computer to make decisions through trial and error, like training a dog with rewards for good behaviour.
- Deep Learning: Using a computer with a brain-like structure to understand complex things, like recognizing faces in photos.
- Transfer Learning: Sharing knowledge between tasks, so if the computer is good at recognizing animals, it can also tell you what’s in your photos.
- Generative Adversarial Networks (GANs): Making the computer create things (like art or photos) by having it compete with itself to improve.
- Self-Supervised Learning: Asking the computer to figure out part of a problem by giving it hints about the solution.
- Evolutionary Algorithms: Solving problems by simulating how things evolve in nature, like how species change over time.
- Ensemble Learning: Combining multiple computers to work together, like a group of friends solving a puzzle.
- Adversarial Learning: Making the computer smarter by having it compete against itself, like a game where it tries to trick itself into being better at a task.
Now, this whole deal with adversarial learning is like teaching a 5-year-old to navigate the maze of a modern grocery store. You know, the ones with many aisles, products you’ve never heard of, and signs that seem to be playing hide-and-seek. Now, the traditional way would be to hand the kid a thick rulebook with directions like, ‘First, go to aisle 3 for cereals, then to aisle 7 for snacks,’ and so on. In Machine learning we don’t do that instead, we say, ‘Imagine you’re hungry, and you need to find something tasty.’ We teach our tot to think for themselves. They can adapt, and if they spot a shortcut to the candy section, they’ll take it! No dull rules, just imaginative models that make things way more exciting.
So ‘how is this not Artificial Intelligence?’ I hear you scream. Remember we said there are many parts to the body and a couple of things don’t make it whole.
Enter the Marketing and PR department
Now marketing is all about catchphrases and skirts on the grey borders of technicalities. I remember an ad for HD glasses, like spectacles you wear to see in High Definition. How? I don’t know. Here is another classic example, Google “What does 360, 480, 720, 1080 mean,” the short answer, it means the number of horizontal lines for your screen.
To call a display with 2160 lines 2K doesn’t look cool on a sticker so why not change the meaning to mean vertical line and all of a sudden it becomes 4K. These are the same folks who go around naming things.
So why is it important to name things correctly?
For one, I am a purist. Just like how car enthusiasts would find it preposterous to call an engine plus 2 wheels a car, I find it weird to call a generative chatbot AI, because that is what it boils down to. Besides my otherwise personal proclivities, a more important and pressing concern; to fight myopia! People, especially the generation that is going to be around for decades to come must know there is more to AI than just a grey screen that miraculously answers all your questions, and yes it really does excel there. We are not going to talk about ChatGPT4, just 3.5 because yes, the older brother does more than just text but we live in the part of the world that the boys and girls in Silicon Valley don’t see as worthwhile for us to have access to.
Now for this part, I am going to involve us in a conversation we were not part of because as far as these developments go, we were but mere spectators. To understand where we are going (the human collective) we need to see where we started, where are we, and, cue drumroll where we are headed.
The credit for laying out these phases of AI advancement goes to some brilliant minds in the field who’ve charted this exciting journey that has been split into 3 stages
- Artificial Narrow Intelligence (ANI, ~1950s-Present): Think of this as the “task-oriented” phase. ANI does specific things really well but struggles with anything beyond its assigned job. It’s like a robot that’s amazing at making coffee but can’t tell you the weather.
- Artificial General Intelligence (AGI, Ongoing): This is a big dream, the “human-like” AI. AGI would be like having a robot buddy who can chat with you about sports, help with math homework, and even make coffee, all in one. But, we’re not there yet, and it’s the tech world’s holy grail.
- Artificial Superintelligence (ASI, The Future): This is where things get sci-fi. ASI would be like a super-genius AI that makes all the decisions, fixes problems, and probably even understands why cats like cardboard boxes so much. It’s the AI boss of everything. Every doomsday AI movie also happens to be based on this.
We certainly were left out on ANI; we definitely can pivot ourselves into AGI and be a part of ASI. Getting left behind is really not the brightest of ideas. The bell has been rung and can’t be unrung, this is us imploring the Zimbabwe government to look into this seriously, our future and even existence depends on it.
How you can get involved
Worry not friend, as impressive as global milestones are, the ship has not sailed yet. And as long as you have breath in your lungs, bonds in your pocket, and redistributed land under your feet, you too can be part of this amazing journey. Now I know, Hollywood paints this picture of 15-year-old whiz kids who can hack the NSA with a calculator and you are left wanting. But fear not, there are many vocations that are available and many niches that will still need to be filled. There are numerous ways for normal human beings who don’t carry laptops to funerals can participate in the AI revolution. Some areas and roles that offer opportunities:
- AI Ethics and Policy: As AI continues to advance, there’s a growing need for individuals who can help shape ethical guidelines and government policies surrounding AI technologies. You can engage in discussions and research related to AI ethics and advocate for responsible AI development.
- AI Education: Educating others about AI is crucial. You can become an AI educator, by conducting workshops, webinars, or writing articles to help people understand the basics of AI and its societal impacts.
- AI Marketing and Content Creation: AI companies need marketing professionals to communicate their technology to the world. You can work on creating content, managing social media, or running AI-related marketing campaigns.
- AI User Experience (UX) and Design: Designers play a crucial role in making AI applications user-friendly and appealing. You can work on designing interfaces for AI-powered apps and ensuring a seamless user experience.
- AI Research and Analysis: Even without a technical background, you can still analyze AI trends, market developments, and research findings. This information can be valuable to businesses and policymakers.
- AI Sales and Business Development: AI companies need sales professionals who can understand the technology and communicate its value to potential clients. If you have strong communication skills, this could be a fit.
- AI Journalism: Covering AI developments in the media is vital for keeping the public informed. You can become an AI journalist, reporting on the latest advancements and their implications.
- AI Consulting and Advisory: Offer consulting services to businesses looking to integrate AI into their operations. You can help them understand how AI can benefit their specific industry.
- AI Project Management: AI projects need effective project managers to ensure they stay on track, meet deadlines, and fulfill their objectives.
- AI Partnerships and Alliances: Facilitate collaborations between AI companies and other organizations, such as universities, nonprofits, or industry partners.
- AI Legal Services: The legal field is evolving to address AI-related issues. Lawyers who specialize in AI can help with issues like intellectual property, data privacy, and regulation compliance.
- AI Startups and Entrepreneurship: If you have a business idea related to AI, you can start your own AI-related venture, even without a deep technical background. You can hire technical talent or collaborate with AI experts.
AI is a multidisciplinary field, and non-technical roles are just as important in ensuring its responsible and ethical development. It’s all about finding the intersection between your skills, interests, and the AI revolution.
Where you do start
Weirdly enough, ChatGPT itself (facepalm) and any other generative chatbot, that is to say, Google Bard. Ask questions, it doesn’t get tired, or bored, ask and re-ask and even add stuff like “in layman’s terms to your questions.” Before you know it, you will have a solid understanding of the basics. Take the next step, there are many good platforms that offer courses from the Introductory to the Advanced stage depending on your level of interest. The one thing that all this asks of you, demands of you, is your drive to learn. You are not going to buy ChatGPT ‘Coke’ and work up with a wealth of knowledge (no disrespect whatsoever meant to Delta Beverage). There is also a budding Data Science community on Twitter, don’t just use it to troll hey.
So, on that note my fellow plebes, peace!
By Guest Author, Lennon I. Emmanuel