AI is having a serious moment. What was once a sci-fi or marketing talking point is now widely available to consumers and businesses (at least in some contexts for some definitions of what counts as artificial intelligence). We haven't got HAL 3000 or Skynet yet, but ChatGPT and Stable Diffusion are at least taking over social media.
Next up? Hopefully business, not the world.
If you've been impressed by these tools, confounded by how they're able to spin long, coherent paragraphs or make photo-realistic images, and just want to know more, I'm about to dig into all that. And if you're scared about what they mean for your career—or excited about what they could do for your business—we'll look at that too.
What is artificial intelligence?
As with what constitutes intelligence in humans, AI is hard to neatly draw a box around. In the broadest possible sense, artificial intelligence refers to computer programs—or machines running computer programs—that can "learn, reason, and act for themselves" as well as "make their own decisions when faced with new situations."
In simple terms, a non-AI computer program is programmed to repeat the same task in the same way every single time. Imagine a robot that's designed to make paper clips by bending a small strip of wire. It takes the few inches of wire and makes the exact same three bends every single time. As long as it keeps being given wire, it will keep bending it into paper clips. Give it a piece of dry spaghetti, however, and it will just snap it. It has no capacity to do anything except bend a strip of wire. It could be reprogrammed, but it can't adapt to a new situation by itself.
AIs, on the other hand, are able to learn and solve more complex and dynamic problems—including ones they haven't faced before. In the race to build a driverless car, no company is trying to teach a computer how to navigate every intersection on every road in the United States. Instead, they're attempting to create computer programs that are able to use a variety of different sensors to assess what's going on around them and react correctly to real-world situations, regardless of if they've ever encountered it before. We're still a long way from a truly driverless car, but it's clear that they can't be created in the same way as regular computer programs. It's just impossible for the programmers to account for every individual case, so you need to build computer systems that are able to adapt.
Of course, you can question if a driverless car would be truly intelligent. The answer is likely a big maybe, but it's certainly more intelligent than a robotic vacuum cleaner for most definitions of intelligence. The real win in AI would be to build an artificial general intelligence (AGI) or strong AI: basically, an AI with human-like intelligence, capable of learning new tasks, conversing and understanding instructions in various forms, and fulfilling all our sci-fi dreams. Again, this is something that's a long way off.
What we have now is sometimes called weak AI, narrow AI, or artificial narrow intelligence (ANI): AIs that are trained to perform specific tasks but aren't able to do everything. This still enables some pretty impressive uses. Apple's Siri and Amazon's Alexa are both fairly simple ANIs, but they can still respond to a wide number of requests.
One last thing: With AI so popular right now, we're likely to see the term thrown around a lot for things where it doesn't really apply. So take it with a grain of salt when you see a brand marketing itself with the concept—do some digging to be sure it's really AI, not just a set of rules. Which brings me to the next point.
How does artificial intelligence work?
Businesses will have the most success adopting AI if they have existing data—like customer queries—to train it with.
Currently, most AIs rely on a process called machine learning to develop the complex algorithms that constitute their ability to act intelligently. There are other areas of AI research—like robotics, computer vision, and natural language processing—that also play a major role in many practical implementations of AI, but the underlying training and development still starts with machine learning.
With machine learning, a computer program is provided with a large training data set—the bigger, the better. Say you want to train a computer to recognize different animals. Your data set could be thousands of photographs of animals paired with a text label describing them. By getting the computer program to crunch through the whole training data set, it could create an algorithm—a series of rules, really—for identifying the different creatures. Instead of a human having to program a list of criteria, the computer program would create its own.
This means that businesses will have the most success adopting AI if they have existing data—like customer queries—to train it with.
Although the specifics get a lot more complicated, structured training using machine learning is at the core of how both GPT-3 (Generative Pre-trained Transformer 3) and Stable Diffusion were developed. GPT-3—the GPT in ChatGPT—was trained on almost 500 billion "tokens" (roughly four characters of text) from books, news articles, and websites around the internet. Stable Diffusion, on the other hand, used the LAOIN-5B dataset, a dataset with 5.85 billion text-image pairs.
From these training datasets, both GPT-3 and Stable Diffusion developed neural networks—complex, many-layered, weighted algorithms modeled after the human brain—that allow them to predict and generate new content based on what they learned from their training data. When you ask ChatGPT a question, it answers by using its neural network to predict what token should come next. When you give Stable Diffusion a prompt, it uses its neural network to modify a set of random noise into an image that matches the text.
Both these neural networks are technically "deep learning algorithms." Although the words are often used interchangeably, a neural network can theoretically be quite simple, while modern AIs rely on deep neural networks that often take into account millions or billions of parameters. This makes their operations murky to end users because the specifics of what they're doing can't easily be deconstructed. These AIs are often black boxes that take an input and return an output—which can cause problems when it comes to biased or otherwise objectionable content.
There are other ways that AIs can be trained as well. AlphaZero taught itself to play chess by playing millions of games against itself. All it knew at the start was the basic rules of the game and the win condition. As it tried different strategies, it learned what worked and what didn't—and even came up with some humans hadn't considered before.
AI for businesses: Examples of artificial intelligence in business
AI-powered systems are increasingly present in the world around you, often in subtle ways. The big, interactive AI use cases like DALL·E 2 and ChatGPT are flashy, but they're far from the most popular ways it's used. Business AI is going to start changing the way we work, so after you're done asking ChatGPT for jokes and posting AI selfies, take a look at these examples of AI in business.
Recommendation algorithms are increasingly developed using machine learning—which is considered a subset of artificial intelligence. Netflix, for example, has published a large amount of research about how it uses machine learning. Similarly, Amazon uses AI-driven recommendation algorithms to suggest new products for people to buy. Even Google Search has a few AI components. But in all of these cases, it operates in the background.
Digital personal assistants
Digital personal assistants like Siri and Alexa operate using conversational AI, the process of simulating the experience of talking with another person. It's hard to label each one an individual AI because they have dozens of different functions all operating using different algorithms. For example, Siri's suggestions for apps to open doesn't use the same neural network as its language recognition or the one that determines what settings you've asked it to set your Philips Hue smart lights to. The overall experience, though, is powered by AI.
Customer care services and chatbots
Customer care services and chatbots increasingly have an AI component. If you've ever called a customer care department that asked you to speak your account number or phone number, it was using a speech recognition AI—though since I have an Irish accent, in my experience, not a very good one. Chatbots can also work significantly better with improved language recognition and sentiment analysis. Rules-based versions that just respond to keywords never feel natural, while AI-based ones can offer a more seamless experience.
GPT-3 can now generate text that appears written by a human. That means it can be used to write blog posts, social media content, emails, and website copy using a service like Jasper. How effective it is depends on the kind of content you're creating and the prompts you give it. Warning: GPT-3 is very good at creating plausible-sounding text, but it also can create plausible nonsense.
AIs can also convert English language text into other languages, and vice versa. Services like Weglot automate the process for businesses that operate in multiple countries. As soon as you add new web copy or a blog post, it gets converted to your target languages. Of course, these kinds of AIs can (and probably will) still miss a bit of nuance, but with human supervision, it can really speed up the process of having an international website.
AI meeting schedulers
AI meeting schedulers are able to automatically book meetings and other appointments based on your requirements and habits. Want only one afternoon meeting per day or need a long break between calls? Instead of having an open calendar where anyone can grab a slot, an AI scheduler can dynamically adjust things as people request chunks of your time. And it learns your preferences, so it can predict when the best times for meetings will be.
While basic auto-correct falls short of being an AI, there's a new generation of AI editors like GitHub Co-pilot that look to make intelligent suggestions in real-time. Instead of just autocorrecting to the most common suggestion, they're able to understand what you're trying to do—and help you do it.
Cybersecurity and fraud detection algorithms also rely on machine learning—and some can be considered implementations of AI. They look for anomalous patterns in huge amounts of data and then act to shut down potential breaches or stolen credit cards.
Business AI can also help you communicate with your clients, partners, and other business contacts. Here's an example of how you can use OpenAI and Zapier to automatically draft emails that reply to messages you receive. Just read them before you click send, please.
More AI examples in business
Those are just some of the more obvious applications of artificial intelligence in business. It could also be used for optimizing work rotations, deciding which clients to prioritize, and even handling things like expenses. Really, the list is endless, and AI will probably make it for us.
The impact of AI on business
There are obviously reasons to be cautious about AI: Tesla's autopilot feature continues to crash cars, Amazon accidentally created a sexist hiring AI, and nearly every major AI product still has some biases because they're trained on content scraped from the internet.
And when it comes to stealing jobs, the growth of AI in business is likely to change things quite a bit. For example, AI content generation tools may not replace humans, but they can certainly increase the speed at which one writer can produce. Similarly, improved chatbots will likely be able to handle more customer support queries and even marketing outreach. It's not that businesses won't need customer care agents, but they'll probably have more of a supervisory role.
Of course, a lot of these tools still haven't been built yet. So for now, AI might be most used as the hot new marketing word.