If you want to understand and use AI, you need to know the very real pros and cons of artificial intelligence.
While AI experts don't agree on many things, they all agree on one thing:
AI technology is going to have huge effects on society and business.
Google CEO Sundar Pichai says AI is "one of the most important things humanity is working on," and is more profound than our development of electricity or fire.
Mark Cuban insists business leaders must understand and embrace AI.
"If any of you are entrepreneurs or in the business world and if you don't know AI, then you're the equivalent of somebody in 1999 saying, 'Yeah, I'm sure this internet thing will be okay, but I don't give a [expletive].' It's the same thing."
And Elon Musk warns "If you're not concerned about AI safety, you should be. Vastly more risk than North Korea."
Here at Marketing AI Institute, we've spent years researching and applying artificial intelligence in digital marketing and sales.
In that time, we've concluded that AI applications will have an overwhelming positive net impact on the world. But we also have a realistic view of the risks and problems that AI presents.
To help you unpack AI, we've compiled the list of the top 20 pros and cons of artificial intelligence that are critically important to understand today.
If you don't know what AI is, this section will give you a quick, non-technical definition before we dive into pros and cons.
AI is when we give machines (software and hardware) human-like abilities.
That means we give machines the ability to mimic human intelligence. We teach machines to see, hear, speak, move, and make decisions.
The difference between AI and traditional technology, however, is that AI has the capacity to make predictions and learn on its own.
Humans design AI to achieve a goal. Then, we train it on data so it learns how best to achieve that goal.
Once it learns well enough, we turn AI loose on new data, which it can then use to achieve goals on its own without direct instruction from a human.
AI does all this by making predictions. It analyzes data, then uses that data to make (hopefully) accurate predictions.
It then learns from every prediction it makes, and optimizes its approach the next time around. As a result, AI gets smarter over time—often on its own.
There are dozens, if not hundreds, of types of artificial intelligence.
In fact, the term "artificial intelligence" is an umbrella term. It describes many different technologies that have this ability to learn and improve on their own.
You may have heard about one or more popular types of artificial intelligence. These include: machine learning, computer vision, image recognition, natural language generation (NLG), natural language processing (NLP), and deep learning, among others.
You don't need to know all of these terms to understand the pros and cons of artificial intelligence...
You just need to understand that AI is different than other technology because of its capacity to learn and improve over time.
AI technology is forecast to have trillions of dollars in economic impact—and there's a reason for that.
AI has serious pros for people and businesses. Which is why, in 2022, major brands like Amazon, Facebook, Google, Microsoft, Apple, and Netflix are powered by AI technologies.
AI excels at automating repetitive, data-driven, and mundane tasks. These tasks are ones that humans spend a lot of time and energy doing-time and energy that can be better used elsewhere.
Today, AI does everything from responding to emails to performing complex manufacturing tasks. This frees up vast amounts of human labor to use on other activities, both at work and in life.
In business, humans aren't very good at consistently and accurately making decisions based on data. We exhibit far too much bias and have too many mental blindspots. Not to mention, we get tired and distracted.
And the same limitations apply to more physical tasks. Even the most proficient human on an assembly line makes many mistakes.
AI reduces human error in many different areas of business and life. That's because AI follows consistent logic and has no feelings that get in the way of analysis. Also, AI doesn't have attention or distraction problems.
This is why you increasingly see AI being used for tasks the need to be error-free, like precision manufacturing or driving assistance.
AI software running in robots can do tasks that humans find dangerous.
Today, AI-powered robots can assist or takeover perilous manufacturing, surveillance, and maintenance work, so that human workers don't have to risk life and limb.
Artificial intelligence has the ability to recognize patterns in big data, then use those patterns to make predictions. In turn, these predictions help you make better decisions.
Google Maps uses AI to predict which routes are optimal, so you can choose the one that gets you to your destination fastest.
Amazon uses AI to predict which product you might like to buy next, which helps you make better, more enjoyable purchasing decisions.
In business, AI can do everything from predicting which equipment in a plant needs maintenance to determining which of your leads are ready to buy.
As one example, eBay used AI to predict which email subject lines customers would open. The predictions were better than those made by human copywriters, and raised average open rates by 15%.
AI also detects patterns in numbers, words, and images better than humans.
By doing this, AI makes your life easier in tons of ways.
You can now securely unlock your phone just by looking at it, since AI detects the unique patterns of your face.
AI finishes your sentences in Gmail because it detects patterns in human writing and knows what comes next.
AI pattern detection even makes it possible to have self-driving cars that identify objects and obstacles in real-time.
We can't recognize patterns like AI can, or at the speed and scale AI can. This is why AI is able to facilitate these types of solutions—solutions that humans can't do or miss entirely.
Here's one incredible example of this fact:
AI was able to find new sources of revenue for a travel business because it found patterns of customer behavior in its advertising data that the company had completely missed.
AI unlocks benefits that traditional software doesn't provide. That's because it can do things traditional software can't and improve on its own, producing compounding benefits over time.
This is projected to have a tremendous effect on the world economy. PwC estimates AI will cause a 14% lift in global GDP possible by 2030, a total contribution of $15.7 trillion to the world economy, thanks to both increased productivity and increased consumption.
This, in turn, will make entire populations richer and better off.
AI is also going to make individual businesses and workers more valuable.
In 2021 alone, Gartner projected AI augmentation would create $2.9 trillion of business value, and save 6.2 billion hours of worker productivity globally.
That's because AI can:
Unlike us, AI has the ability to work on problems and learn from their solutions all day, every day.
In fact, the longer and more often AI tackles problems, the smarter and smarter it gets at solving those problems.
To experience all the pros of AI, you need to have a clear, realistic understanding of its cons.
AI is only as good as the amount and quality of data it has.
If the latest AI app on your phone doesn't have enough data, it will produce bad results.
The same is true in business. Many companies need a minimum amount of data to get started using custom AI models or some AI tools.
And that data has to be high-quality and clean. For some, significant work and investment is required to get internal data AI-ready.
The exceptions are AI tools that use third-party datasets. Many of these already exist, and use either a proprietary dataset the vendor owns or collect data from online sources, then apply proprietary algorithms to it.
Understanding what data you need for an AI solution is a critical step. And it's not always easy or fast to figure this out.
Often, trained data scientists are needed either full-time or on a consulting basis to clean and organize data for use with AI.
When AI goes wrong, it can go really wrong.
AI is also not going to become self-aware and take over the world. That's science fiction. But it can still make big errors or bad decisions. When it does, negative effects happen at scale.
AI systems are deployed at scale across millions of devices. If AI starts making bad or harmful decisions, it could hurt millions of people physically or financially.
A good example of this is self-driving cars. If the AI in charge of a brand of self-driving cars has a flaw, that flaw could show up in thousands or millions of vehicles.
Some AI systems are what we call "black box" systems. The user has little to no understanding of how the AI makes decisions.
This isn't a dealbreaker when the system works well.
If you are getting good value out of an AI tool, you typically don't care how it works.
But if the AI makes a poor prediction, takes a damaging action, or makes a mistake, it can be hard or impossible to diagnose what went wrong.
This makes it hard for people and companies to trust AI with important decisions.
It's also why there's been a push for more explainability in AI tools.
AI uses data to make decisions and predictions. That data might contain conscious or unconscious bias. If it does, then an AI system could make decisions that discriminate against certain groups or types of people.
For instance, AI systems can use data that is inherently flawed, which then causes bias and/or discrimination.
In one example, an Amazon hiring algorithm developed bias towards female job candidates thanks to the data it was using.
The data was comprised mostly of resumes from men, so the machine mistakenly assumed that one quality of an ideal job candidate was being a male.
It's impossible to predict with a high degree of accuracy how many jobs AI will take. And, we think AI will create and enhance far more jobs than it eliminates.
However, the danger is always present that AI will get good enough at enough tasks to cause widespread job loss and long-term unemployment.
The world of artificial intelligence is filled with hype, buzz, and larger-than-life claims.
While there's no doubt AI is going to transform the economy, it still has limits and specific use cases. It can't do everything. And it may not always be the right tool for a business.
AI excels at performing narrow tasks extremely well, on its own, at scale.
But the level of advancement in different fields of AI is uneven.
Some areas of AI, like language generation and computer vision, have progressed significantly. Other areas are still just scratching the surface of what's possible.
In reality, AI can do many narrow tasks much better than humans, but it's still math, not magic.
Companies and individuals need to temper expectations about the technology's capabilities, and take the time to fully understand what AI can and can't do, so they can actually select, pilot, and scale the technology effectively.
As a consumer, you benefit from AI in everyday life. You rarely have to purchase AI tools on your own. Instead, companies use AI to provide better, more profitable consumer experiences that end up serving you.
However, if you're a business, some types of AI can be prohibitively expensive.
Make no mistake, there are AI tools that are affordable for every business. But many of the most advanced systems or custom machine learning models can cost a large amount of money to implement or develop.
There's no question AI will take over some of the tasks done by your average human worker...
But AI can't do everything.
In fact, there are some broad rules to consider when asking yourself if AI can do your job better than you.
We don't think it's likely that AI is going to replace your average human being.
While AI may automate jobs with the qualities we listed above, it's increasingly unlikely that humans only do extremely narrow, repetitive, or only data-driven tasks.
Your typical knowledge worker today wears many hats, and performs many creative and strategic tasks that AI just can't do.
We don't foresee AI replacing a critical mass of human workers. In fact, we generally predict AI will enhance and augment our work.
You see, a lot of tasks that AI can do better than humans are tasks that humans weren't that good at to begin with.
Few of us can even do pattern recognition or make fully data-backed predictions well.
And none of us can perform those tasks at scale. When we try to, the result is either wrong or imperfect and it usually takes a lot of time, energy, and money to produce.
By letting AI do what it's good at, we free ourselves up to do what we're good at. And what we're good at is usually higher value work like creative output and strategic decision-making.
We don't think you should fear AI, but you should be wary of another barrier to AI adoption...
You should, to the point made by Elon Musk in this article's introduction, fear the possibility of AI going wrong at scale. As outlined in the cons section, AI can be misused or create negative outcomes.
But we're not seeing a lot of businesses fear AI just because of what it can do.
The COVID-19 pandemic has accelerated AI-powered digital transformation across businesses.
Research from McKinsey cites that 25 percent of almost 2,400 business leaders surveyed said they increased AI adoption due to the pandemic.
The 2021 State of Marketing AI Report (in partnership with Drift), we found that among marketing leaders, fear wasn't a major barrier to AI adoption.
The majority of marketers we surveyed (56%) believe AI will create more jobs than it eliminates over the next decade. Only 16% said fear was a contributing factor to their company's lack of adoption.