What Is Artificial Intelligence? What AI Means for Our Everyday Lives

Last week, I was running late for a meeting with an important new client. I thought I gave myself enough time, but about halfway through the ride, my AI scheduling and calendar app notified me of heavy traffic thanks to an accident.

Within seconds, it asked Siri for alternative directions. That was one of the big moments where I thought, “Wow, AI just saved my butt.” It reminded me that AI isn’t a concept in some science fiction book. It’s here to stay.

Artificial intelligence, or AI for short, involves computer science and its applications for building software and machines that mimic human intelligence, most notably learning, reasoning, and making decisions like humans.

With complex algorithms and language learning models, the world already has various examples of artificial intelligence that can perform tasks and adapt to new information. I’m here to explain it in the most basic terms possible — but also give you a dive into the more complex topics surrounding AI.

Imagine someone going to school to learn to be a dentist. They’d study the newest literature, look at charts and pictures, and eventually be able to identify things like cavities and receding gum lines.

Artificial intelligence works similarly, except we’re giving machines the information (like images and documents) on certain topics and teaching them to process that information. In the end, the goal is for the machine to make decisions with that information.

That’s where it becomes artificial intelligence. Humans learn from experiences and use that knowledge to complete tasks. AI learns from training (experience), then uses databases and algorithms to simulate knowledge processing and decision-making.

Let me break down AI into several key components. I find it helpful to compare AI to how a chef manages her craft.

Key components:

• Learning: An AI system “learns” by accepting data like images, documents, and search results. It analyzes and stores this data in its database. A chef learns similarly, reading cookbooks, recipes, and menus, while also taking classes, eating food, and practicing to store all that new information in her brain.
• Reasoning: The AI system can then make decisions that are both logical and based on the data it has already consumed. A chef does the same thing, choosing the right ingredients based on her previously consumed knowledge.
• Problem-solving: This involves finding solutions to overly complicated problems, which is still hard for AI, though it is improving. A chef, for instance, might fix a burnt dish with creative solutions instead of starting over.
• Perception: AI should not only process elements from the real world but understand them for further processing. It’s similar to how a chef might perceive if food is done.
• Language understanding: This is similar to perception but has more to do with understanding and responding to human communications as opposed to other elements in the real world. For a chef, they must respond to praise or criticism sent back from diners.

With these key components, we’re starting to see a growing role of AI in our everyday lives. I see it all the time; in society, in the economy, and in technology. AI does everything from helping Netflix recommend my next show to helping doctors diagnose diseases.

The easiest way I’ve been able to comprehend the capabilities of AI is by picturing a ladder. Since AI is currently in its infancy, that means we’re on the lower few rungs of the artificial intelligence ladder. As the years pass, however, developers and researchers work to climb higher on the ladder with more advanced capabilities.

Based on Capabilities

In this section, I’ll break down AI by its capabilities. Imagine you’ve walked into a storefront selling three models of robots, all of which are marketed as AI assistants. Here are the models based on capability:

• Narrow AI (Weak AI): This level of AI is like a worker who’s highly trained in one skill alone. It’s the type of AI we have today in voice assistants like Siri and Alexa. It can tell you the weather and find specific products but has limited abilities to give you personalized fashion advice or analysis on the music you choose. In short, it’s limited to its programming.
• General AI (Strong AI): Here’s what AI researchers want to achieve someday with the technology. We’ve only seen it in science fiction movies, though. It’s an AI that functions much like a regular human employee, one that can use reasoning, take context into play, and have meaningful conversations about anything.
• Superintelligent AI: This is the level of capability where AI surpasses the intelligence of humans. I’m talking about a genius-level machine that can not only out think and outwork every human in the world but make more thoughtful decisions based on reason. This is the theoretical level of AI that many people fear, and lots of researchers wonder if it’s even possible. If it is possible, we’re decades, or even centuries, away from development.

We’re currently in a world with Narrow AI. That’s a capability level that many people have become comfortable with, but there’s more in store. Researchers want a more honed version of Narrow AI that turns into General AI. Only time will tell if we’ll see it soon.

Based on Functionality

The capabilities of AI look into its basic evolution and where we might end up in the future. The functionality of AI, however, tells a different story. It tells us exactly how AI learns and reacts, and which levels of functionality are appropriate for certain applications.

• Reactive Machines: These are basic AI systems without memory. Deep Blue, the supercomputer chess bot designed to beat the world’s best human chess players, was a reactive machine. It sees your move, looks at the board, then makes a reflex action based on what’s currently going on in the game. Reactive machines stay in the present, never logging or learning from their past actions or experiences.
• Limited Memory: Self-driving cars use Limited Memory AIs. That’s because they not only have to respond to current road conditions (which would only require a reactive feature) but they must use past experiences to make better decisions and anticipate problems. This helps the AI learn, like how it might identify your narrow driveway that requires slower, more careful parking in the future.
• Theory of Mind: This type of functionality is still only a theory, but many researchers are working on projects with Theory of Mind AI. They’re advanced systems that actually comprehend emotions from humans. For instance, a virtual assistant could schedule a call with someone but only make it short if you sounded annoyed or angry on the previous call.
• Self-Aware AI: Here’s a feature of AI that’s purely speculative. It involves creating a machine with AI that mimics human consciousness. As such, a Self-Aware AI could intentionally lie, experience emotions, and even show empathy.

I can’t imagine ever seeing a Self-Aware AI in my lifetime, but it would be interesting to experience a Theory of Mind AI with the potential to comprehend my emotions.

One flaw of AI, in my opinion, is its inability to read context and emotional cues. That’s a huge part of being human. So, until then, I can’t say that AI is truly mimicking human behavior.

I find it helpful to think of AI as a human body. As I discuss the core components of artificial intelligence, I’ll explain how it relates to a certain part of the human body, like the brain, mouth, or eyes.

Machine Learning (ML)

I view machine learning as the brain of artificial intelligence. You’ve probably heard about machine learning and perhaps assumed it’s the same thing as AI. That’s not the case, however. It’s only one component of AI — the part that learns information from sources and stores that information for later use.

You’ll see machine learning in several applications:

• Fraud detection for credit cards and online purchases
• Recommendations (like for Netflix or Hulu)
• Spam detection and removal

There are three types of learning in machine learning: supervised, unsupervised, and reinforced. Supervised machine learning works with direct guides who tell the machines what to learn. Unsupervised learning happens by letting machines explore and learn topics on their own. Finally, reinforced learning promotes good decisions with rewards.

Natural Language Processing (NLP)

I see NLP, or natural language processing, as the ears of the AI operation. Closely linked to the brain, NLP helps machines absorb and process details like semantics, syntax, and ambiguity.

Here’s where it comes into play:

• To understand what the response would be to weather or location-based questions. For instance, “what’s the weather like?” could produce a range of responses, but the AI should know that you’re only worried about the weather in your town, not elsewhere.
• When translation apps handle complex sentences that change based on cultural nuances, rather than just providing word-for-word translations.
• When chatbots in online stores reply to customer questions.

People speak in unique tones, and they may make statements with nuances or sarcasm. The goal with natural language processing is to catch all that for the best response.

Computer Vision

As you may have already assumed, computer vision is like the eyes of any AI operation. It shows machines how to view the world as humans do. This way, AI can look at faces, pictures, and other visuals to make decisions.

Computer vision allows for:

• Facial recognition on devices.
• Pedestrian and obstacle detection in self-driving cars.
• Tumor spotting for doctors.
• Package detection for smart doorbells

Computer vision not only provides a way for machines to “view” content, but it helps with processing those visuals, too. I consider it one of the most important parts of AI in today’s technology, since we see it used in so many products.

Robotics

Robotics function as the joints and appendages of AI. It’s the link between an AI’s brain and physical activity. Otherwise, AI can only provide thoughts through writing or speaking.

Here’s where you might see robotics in use:

• Deep ocean or space exploration robots that venture into perilous environments.
• Manufacturing robots that build items like cars and smartphones with precision unmatched by humans.
• Surgical robots that help doctors with procedures.

The vast majority of robotics are far from what you’d see in a science fiction movie. That’s partially because we’re still at the level of Narrow AI with reactive machines and limited memories. And it’s incredibly difficult to replicate the complexities of human joints, muscles, and bones.

Neural Networks and Deep Learning

Much like the constantly firing neurons in our brains, researchers have attempted to build artificial neural networks that connect for a robust architecture of links. Although in its infancy, neural designs could lead to some advanced deep learning options.

Here’s where neural networks could take us:

• Image recognition that’s better at identifying items than the average human.
• Language translation that improves over time and offers greater sensitivity to things like colloquialisms and emotions.
• Voice recognition that learns about you through your smart speakers.

The idea is to link large swaths of data with artificial neural networks, allowing AI to rapidly access that data and simulate the way a human brain might scour its own memory.

I think it’s essential to understand the core components and capabilities of AI long before you learn about how it’s used in our world. This way, you have a deeper understanding of how it could affect your personal and professional life.

But once you’ve understood all that, I’d like to explain the real-world applications of artificial intelligence. There’s a good chance it affects you in some way.

You’ve probably heard arguments for and against artificial intelligence. Perhaps you’ve formed your own opinions based on movies, journalists, or even how AI has already affected your industry.

There’s no doubt: AI has extreme ethical and social implications.

As a writer, I’ve worried about job displacement and AI tools simply scraping the internet to steal from thousands of writers like myself. Then there’s bias, privacy concerns, autonomy, and other issues I’ll cover below.

Bias and Fairness

I’ve read about several instances of bias in AI technology. One podcast I listened to discussed preferences in whom a self-driving car protects.

Then there’s facial recognition software, which has struggled with identifying certain shades of skin.

I also read about how an AI recruitment tool got scrapped once someone discovered it had an extreme bias against women (thanks to its training on a history of male-centric hiring patterns).

Privacy Concerns

Data sharing happens constantly in the world of AI. When I speak to Siri, my data goes to a server somewhere. Almost all tech companies are guilty of problematic data collection policies.

There’s also surveillance. If every microphone and camera records everything we do, isn’t that a problem?

And what about how those machines process the data? We have no idea where our images and words go and how they’re being used.

Job Displacement

The jury is still out on whether AI will create more jobs than it will destroy. And it’s difficult to analyze the true results of such technology because some new developments can claim they produce “more jobs” but, when you look into it, those jobs are a lot less desirable and lower paying.

So, AI could replace many people in industries like media, health, manufacturing, and logistics.

It’s obviously important for workers to stay ahead of the conversation and adapt, but it’s equally wise for people to stand together and push back against so-called “disrupting” technologies that simply shift wealth from thousands of workers to a small group of people.

Autonomy and Control

Autonomous systems always carry the risk of acting against human interests. Right now, there’s the potential for malicious developers programming harmful elements into AI systems.

Or, as I’ve seen many times in the news, researchers unintentionally create an AI system that acts against human interests — and that’s because there’s a flaw in the training materials.

And as the machines and systems behind AI become more sophisticated, you should know about the threat to autonomy. In short, this means an AI system could act without a prompt from a human.

It’s like the rogue AIs we’ve seen in movies (like HAL from 2001: A Space Odyssey). The AI becomes self-sufficient. It no longer needs humans to operate, and for whatever reason (spite, perhaps?), the AI acts against what we would want as humans.

AI in Decision-Making

Consider this: an AI system decides how a doctor proceeds in a surgery. But the AI was wrong, so that person died.

Or maybe an autonomous vehicle drives off a bridge because it decides in a split second that the mother and child it’s trying to avoid on the street are more valuable than the single man driving the vehicle.

These aren’t just theoretical questions. They’re legitimate queries on how we intend to let AI control decision-making in the real world.

There’s more to being an AI developer than simply hammering out enough clever coding to make it work. AI demands moral researchers willing to abandon promising ideas for the sake of societal values.

Those in the AI space must have morals and the willingness to confront some of the most complex challenges that span virtually every industry and discipline in the world. I’ll talk about the challenges in AI development below.

Technical Challenges

Beyond the moral issues of artificial intelligence, there are constant technical challenges. I, for one, have often wondered what AI developers plan to do once the entire internet has been scraped of its content and all the writers put out of work.

I’d imagine a constant regurgitation of material until we have an internet filled with poorly made, inaccurate content.

Here are my main technical challenges with AI:

• Computing power: Advanced AI models demand massive amounts of energy and computing power. Sustainability experts already say how damaging AI is to the environment, so we need to figure out ways to get higher levels of computing power for less money and without destroying the environment.
• Data quality and quantity: Every AI model needs to start with good data, even if it’s not from the internet. Otherwise, AI won’t function properly.
• Scalability: The goal is to keep scaling AI systems so they can handle the complexities of the real world, but every growth step is harder and harder.

Data quality is the first big challenge. Then there are computational limitations. I’m uncertain we’ll ever totally struggle to scale AI, but developers may have trouble with scaling if they’re limited by computing power and data shortages.

Regulatory and Legal Issues

We’re in strange waters with AI when it comes to legal issues and regulation. Who owns AI-generated art? I’d argue it’s a mashup of artwork stolen from hundreds of real artists, but that’s just me.

Beyond that, who’s liable if AI makes a mistake? We’re already seeing that police officers do not know how to respond if they pull over an autonomous vehicle. It seems comical in the news, but these issues need solutions. Otherwise, we’ll live in a world with no accountability.

Interdisciplinary Collaboration

Speaking of regulations and legal issues, I think it’s essential that all AI projects receive input from experts in the fields of psychology, sociology, and ethics.

Sharing a discussion with experts from these disciplines can help tech companies align their vision and AI solutions with actual human needs.

It makes me cringe when a tech CEO stands in front of a crowd to unveil their new AI product because, more often than not, they used no input from people who truly understand human behavior.

The only way to build well-rounded, safe AI solutions is to get input from more than a small room of tech developers trying to make money.

Energy Consumption

Even small-scale AI models consume an insane amount of energy. And the large models will continue to face energy issues unless they figure out more eco-friendly ways to train, produce, and manage AI models.

I’ve received this question many times. What will AI bring in the future? Can we expect it to take over the world? Will it take all our jobs? While I lack the ability to see the future, there are some worthwhile trends and ideas to consider.

Emerging Trends

My favorite emerging trends in AI involve those that inherently help the world, instead of another tech company trying to make a billion dollars without thinking about the impact of their work.

Here’s what to look forward to:

• AI for climate change: I’d like to see more of a focus on using AI to tackle environmental issues.
• Explainable AI: This is a shift from the basic AI systems that simply provide feedback. Instead, this type of AI can explain the “why” behind its feedback and decisions.
• Collaboration between humans and AIs: I like this trend because it strives to balance human intuition and AI capabilities instead of just constantly pushing for more AI.

I think everyone should keep an eye on how developers incorporate that human-AI collaboration I spoke about above. At what point have we created enough AI to complement our current human intelligence? There’s no way we can just keep building more advanced systems and still be providing value to humanity.

Speculations and Predictions

My prediction is that we’re still centuries away from artificial general intelligence (AGI) thanks to our lack of knowledge surrounding intelligence and consciousness.

Others, however, predict we’re closer to a few decades away from that conclusion. Singularity — when AI surpasses the intelligence of humans — continues to divide experts. Some say it could happen around 2050, but I think that’s just people wanting to see it in their lifetimes.

Role of Governments and Organizations

Countless countries are working on AI strategies. This means some healthy competition with a large dose of noncooperation. In short, countries need to understand that international cooperation is necessary for the most ethical and effective movement forward with AI technology.

Balancing Innovation and Ethics

Many AI developments have provided advancements in technology, but the truly valuable AI developments will have more than that. I want to see a balance of sustainability and equity packed into every major advancement in artificial intelligence. This way, AI developments benefit all of humanity instead of just lining the pockets of a select few.

You’ve heard the rhetoric: AI is here to stay. It isn’t stopping for anyone. You better learn to adapt. Yes, AI is a huge deal, but most people who act like you absolutely must adapt are probably making money from AI or just regurgitating what they heard on the news.

Here’s the thing. AI needs humans, and the whole point of artificial intelligence is to improve our lives.

Therefore, we need all AI development to include input from experts in human psychology and ethics, and to always put humanity’s best interests before quick profits.

And from your perspective, you’re best off learning as much about AI as possible and speaking up or voting on it when your time comes.