Artificial intelligence, cognitive computing, machine learning, data mining, natural language processing: The terminology surrounding the efforts to enhance how businesses access, sort and analyze the vast quantities of content being created every day is almost as vast as the content itself — and no less confusing. The words are often used interchangeably, and while some have clear definitions, others do not.
Claims that artificial intelligence (AI) will transform health care, law and “business” as a whole are rampant, but with the term left undefined, it is difficult to sort what is blue sky and what is real opportunity.
With everything from smart homes to self-driving cars to chatbots being labeled as artificial intelligence, the conversation should really start with a clear working definition of what AI is and what it’s not.
What Is AI Exactly?
Artificial intelligence, in the broadest sense, is the act of a machine solving a problem similar to the way a human would.
General AI doesn’t exist and likely never will. Instead, what is considered applied AI today is often the act of using software tools to help automate and speed up known processes. However, there are clear limits to the extent of this automation, most notably the problem of context and general knowledge.
For example, if a health care contract states, “All dental treatments are covered,” a computer would not be able to analyze this as an answer to the question “Are root canals covered?” without knowing that “root canal” is a form of dental treatment. The answer to that question requires information that is not contained in the contract but in the knowledge of the person interpreting it. This knowledge cannot be acquired by a computer system without extensive human support.
Today’s most useful AI tools are closer to what is being called “aided intelligence” (a.I.) rather than true artificial intelligence. With aided intelligence, machine-driven approaches support a workflow that is still executed by humans. The individual is still responsible and accountable for applying their vast body of human experience and intelligence to the process — just with a much smaller amount of material to work through. Tasks can be completed more efficiently, with greater accuracy, in a way that is more pleasurable due to the fact that the computer takes care of the routine elements.
For example, while AI will not be replacing lawyers any time soon, a.I. can significantly speed up an eDiscovery or M&A process and increase the quality of the results. Rather than a lawyer reviewing and assessing all the contracts and then bringing in experts to look at particular compliance clauses, for example, aided intelligence automates the initial contract analysis, finds all possible compliance clauses, deduplicates and compares them and presents the results to the expert, thus enabling the allocation of more resources. […]