Artificial intelligence (AI) has the potential to transform businesses forever, equipping them with advanced technological processes that can save time, save money, and improve productivity. But there are some fundamental limitations to what AI can do and fundamental misconceptions about how it works.
Are business leaders overhyping, overestimating, or misinterpreting AI? And if so, how can we address that while still maximizing AI value?
The Nature and Power of AI
Before delving deeper into the applications and types of AI in business, it’s crucial to address a fundamental question: ‘What is Artificial Intelligence (AI)?‘ At its core, AI is any artificially constructed system in computing that’s designed to replicate processes that resemble human intelligence. This includes the ability to learn, adapt, reason, and perhaps even understand. AI systems, primarily based on neural networks and machine learning, strive to mimic the complex, interconnected workings of the human brain.
Make no mistake: AI for business is incredibly powerful and has practically unlimited potential, at least in the long term. But it’s not a magic wand, and not all types of AI are the same.
AI is any artificially constructed system in computing that’s meant to replicate processes that resemble intelligence in some key way. In many cases, AI depends on neural networks, which use interconnected sequences of nodes to replicate the complex, connected systems of the neurons in our brains, in combination with machine learning (ML), which allows a system to continuously learn, adapt, and improve at a given task.
There are also different types of AI, based on its breadth and capabilities:
- ANI. Artificial narrow intelligence is the most common type of modern AI, and the only one to which we have practical access. This type of AI system is extremely good at one common narrowly defined task; for example, it might be good at processing data, translating text, or replicating human language in such a way that it can generate new content. There are fundamental limitations to what you can do with narrow AI, but it tends to be extremely good at one thing.
- AGI. Artificial general intelligence is broader in scope, capable of applying its intelligence to a wide range of different disciplines. Right now, AGI is merely theoretical. While ChatGPT and other conversational AI systems may seem generally intelligent, this is more of a parlor trick than anything.
- ASI. Artificial superintelligence is a vision for a distant future in which AI becomes not only intelligent beyond the scope of human intelligence, but also, perhaps, self-aware. While certainly fascinating, this type of intelligence is so far beyond us that it’s not worth discussing in the context of business AI applications.
The Modern Hype for AI
Today’s business leaders are extremely excited about the potential for AI, and understandably so. However, it’s possible that the hype has eclipsed the current capabilities of AI. As we’ll see in the next section, narrow intelligence is the only type of AI to which we have access, and it’s fundamentally limited in what it can do for businesses.
AI tools have reached a level of sophistication that would have been unthinkable even a few years ago. Business leaders can use AI to automate most predictable tasks, analyze data, and even generate some types of content. It has the power to save money, increase productivity, and even guide businesses in new directions with better, more data-driven insights. However, some business leaders have overestimated or overhyped AI; while it’s truly an amazing tool, it can’t do everything, and even its most appropriate applications require some human guidance and intervention.
Key Limitations of Modern AI
Most businesses are right to pursue the development, exploration, and integration of AI, but there are some limitations to how much AI can be part of your business strategy.
- Context. AI often doesn’t have the context that it needs to provide better guidance or produce better content. It can tell you a lot about a set of data that you give it, but it doesn’t have an understanding of the world that allows it to interpret that data set the same way a human being would.
- “Common sense.” If you’ve played around with language-generating AI, you’ve probably already noticed it doesn’t have much “common sense.” It may be able to produce prose on par with the quality of prose generated by a human writer, but he can’t do basic math or form reasonable deductions about a set of premises.
- Transparency. Modern AI systems follow processes inside a kind of “black box.” Researchers and scientists don’t know exactly what’s happening “under the hood,” meaning they’re not perfectly aware of the patterns and obstacles associated with each system. This makes it hard to understand exactly what’s going on and use that information to produce better systems in the future.
- Bias. Many critics have pointed out that AI is inherently developed with both consciously recognizable and unrecognizable biases. It interprets data in a skewed way, no matter what valiant efforts are undertaken by its developers. Depending on the application, this effect may be significant or negligible.
- Creativity/ingenuity. While AI can simulate creative production, and it can respond to creative-minded individuals issuing prompts to it, AI itself doesn’t have any real creativity or ingenuity. It’s not going to come up with novel ideas in the same way a human being can.
- Emotion/personality. AI systems don’t feel emotions, and there’s debate as to whether any artificial system will ever be able to feel emotions. The same is true for having a personality. There are some applications for which emotions and personality truly matter, and it’s unlikely that AI will be able to convincingly simulate emotion or personality in the near future.
Hype and the Future
Nobody knows exactly where AI goes from here. In a decade, we may be under the rule of a sophisticated AI overlord, or we may be only a few baby steps ahead of where we are now. Either way, it does seem appropriate to suggest that hype has reached a high point; most people understand and are excited about the potential of AI, but some people have overestimated or oversold it.
The best way to harness the power of AI for business is to appreciate it for what it is, use it appropriately, and keep your expectations in check as AI continues to develop.