Artificial intelligence (AI) and machine learning (ML) are two terms that seem to be thrown around interchangeably. Conflating these ideas is understandable; individuals discussing machine learning technologies and companies providing machine learning services often swap these terms. AI and ML, however, are not synonymous. In brief, one term is a subset of the other, but to accurately understand what technologies exists, and which ones companies truly offer, a more nuanced distinction between AI and ML is required.

 

Historically, artificial intelligence refers to simulating human intelligence or replicating human behavior with machines. In scope, this is quite broad. Are we stating that AI needs to be indistinguishable from humans across general cognitive tasks? Are we stating that AI only needs to exhibit some type intelligence toward a specific designated task? The latter type, called weak or narrow AI, is often what companies mean when describing their technologies and services that utilize AI. This ranges from hardcoded rulesets for medical diagnoses to autonomous cars and mobile assistants powered by machine learning. Important to note in these examples is how a medical diagnosis ruleset is not actually powered by machine learning, even though it is AI. Machine learning only describes quantitative processes that use computing systems to learn patterns in data; learning means that those patterns are discovered as a direct result of the data being observed. Given that a hardcoded ruleset is not learning, it is therefore not machine learning. Since it simulates intelligence though, it is AI. This simulated intelligence, however, clearly is not at the level of general human cognition. Since our earlier definition of AI lacking general human cognition was called weak AI, it makes sense to call AI that captures human cognitive capabilities as strong AI. Strong AI would have the ability to react appropriately to unfamiliar tasks. And in the most powerful sense, it may even experience consciousness and self-awareness. This type of AI, however, remains unrealized at this point in time.

 

 

Given that artificial intelligence has multiple interpretations, why do people and companies continue to use the term? Part of this is due to misunderstanding, but another reason relates to marketing. AI sounds revolutionary. A company advertising themselves as sophisticated as possible may choose the label AI, even if somewhat misleading. To avoid any confusion on this front, Komodo Technologies chooses to describe its services as Data Science Solutions. Data Science Solutions accurately represents the capabilities of companies specializing in how to use data to answer questions. And to reach these answers, we can employ machine learning methodologies as a powerful toolset. So if ever faced with a company offering AI solutions, remember to ask for a more nuanced description of the services they offer.