Humans Powered By AI Is A Potent Force

Augmented Intelligence

Gartner defines Augmented Intelligence as a design pattern for a human-centred partnership model of people and artificial intelligence (AI) working together to enhance cognitive performance, including learning, decision making and new experiences. As the name suggests, augmented intelligence uses machines to support humans and not replace them. 

The idea is to use human cognitive faculties and the machine’s speed and accuracy to improve outcomes. Gartner’s Svetlana Sicular, a thought leader in the subject, believes that ‘human plus AI’ is smarter than either by themselves. She says “Augmented Intelligence is all about people taking advantage of AI. As AI technology evolves, augmented intelligence's combined human and AI capabilities will deliver the greatest benefits to enterprises”. Alex Bates, author of Augmented Mind: AI, Humans and the Superhuman Revolution, also believes that augmenting the capabilities of humans and not replacing them is where the real opportunity of AI lies. Augmented intelligence focuses on supporting humans and not replacing them, which is essentially what artificial intelligence aims for.

Way back in 1956, William Ross Ashby used the term Intelligence Amplification in his book Introduction to Cybernetics and this probably is the earliest precursor to the term augmented intelligence. In the book, he argues that just as sound, power and other such attributes can be amplified, so also intellectual power can be amplified, in other words ‒ augmented. (For ease of readability, the term IA will be used in reference to augmented intelligence in this article.)

*Use cases of IA

These days, cars are the perfect example of AI at work. In self-drive mode, the car takes full responsibility of staying in the lane, maintaining a safe distance from the vehicle in front, and going at the right speed. Driver-assist mode provides lane change signals, collision aversion alerts, etc. The former is artificial intelligence, while the latter is augmented intelligence.

Political parties worldwide could not be left behind and have jumped onto the augmented intelligence bandwagon.  Politicians try to influence and sway voting behaviour using data from past elections, demographics, and current affairs. It is believed that automated social media bots were used in the 2016 US Presidential Elections to increase pro-Trump hashtags on Twitter. This is classical augmented intelligence, wherein the target audience, in this case, voters, is presented with data patterns, and the decision-making is left to them.

Augmented intelligence algorithms have applications in healthcare by providing inputs based on various data parameters of patients to medical professionals, helping in the accuracy of diagnosis, personalised prognosis and line of treatment, among others. The expected outcome is reduced human diagnosis errors and improved clinical outcomes. It is also helping in the early detection of life-threatening diseases like cancer, heart, liver, Parkinson’s, Alzheimer’s, etc.

In these times of a connected world, where our personal life is out there in the cyber domain, security is a concern for individuals and governments worldwide. Augmented intelligence systems provide humans with information on possible vulnerabilities, threats and what-if options to help them make informed decisions based on the context.

There is not much information in the public domain for obvious reasons, but these systems have extensive usage in warfare and are in use by the armed forces of many countries. Additionally, we can find many augmented intelligence applications in sectors like financial services, telecommunications, manufacturing and retail.

*AI vs IA

While these acronyms are used interchangeably, there are obvious distinctions between them, and it is important to highlight the key difference between the two AI streams.

AI systems are expected to operate independently and handle tasks that humans perform without inputs from them, whereas IA systems are primarily decision support applications that present patterns based on analysis of humungous volumes of data and then let humans do the decision-making. Artificial Intelligence focuses on the job of replacing humans using fully automated systems. Augmented Intelligence (IA) is focused on supplementing and complementing human functions. The question to be asked is why do we need IA when AI can do things faster, eliminate human errors and thereby create higher efficiency and productivity?

Let’s look at things AI cannot do to answer this question. First and foremost, they cannot create or invent. They do not understand morals or ethics. They are emotionless, have no feelings and cannot handle tasks where human emotion plays a decisive role. Interestingly, AI cannot write software. In his classic on software engineering, The Mythical Man Month, Fredrick Brooks argues that the reason for this is that it doesn’t have the human capability of understanding and comprehension and converting it into rules.

So, which is better, AI or IA? This is a question that does not have a binary answer. In their wonderful article in the Harvard Business Review, David de Cremer and Gary Kasparov wrote: “People and AI both bring different abilities and strengths to the table. The real question is, how can human intelligence work with artificial intelligence to produce augmented intelligence”. Grandmaster Kasparov lost to IBM’s supercomputer named Deep Blue in 1997. This made him rethink the game of chess, and after experimenting in different game situations, he came up with this interesting insight – “Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.” Therein lies the answer to the question; need we say anything more?

An IBM report, Augmented Intelligence Requires Human Direction, references a clinical study that offers another excellent example. In the study, an AI system had a 7.5 per cent error rate when detecting lymph node cancer cells, while human pathologists had a 3.5 per cent error rate. When input from both the AI system and pathologists was combined, the error rate dropped to 0.5 per cent.

Gartner’s AI business value forecast highlights decision support/augmentation as the largest type of AI by business value-add with the fewest early barriers to adoption. By 2030, decision support/augmentation will surpass all other types of AI initiatives to account for 44 per cent of the global AI-derived business value.

In conclusion, one can safely predict that AI will not replace humans in the foreseeable future entirely.

͟͟(This is the concluding part of a two-part article aimed at throwing light on the intriguing field of Artificial Intelligence.)

Jayesh Shah

Guest Author Jayesh Shah