Artificial Intelligence Is Still Far From Intelligent

By Thomas LaRock, Head Geek, SolarWinds

Artificial intelligence (AI) plays a vital role in the digital transformation journey of most businesses and governments.

From smart devices that executes voice commands to self-activating smart lights, AI is intricately integrated into the most unnoticeable aspects of our daily routine.

According to a study by IDC, AI and cognitive systems are the foundations of digital transformation initiatives in the region, and around 70% of enterprises will use AI Services by 2021.

Although there have been fears and scepticism that AI may result in many problems, such as the misuse of the technology and elimination of jobs, there are still significant opportunities driving its development.

From eliminating healthcare inefficiencies to empowering aging economies, AI promises a solution to these problems that are plaguing different industries.

In fact, as a result of this demand for integrated AI solutions, the spending on cognitive and artificial intelligence (AI) in Asia Pacific (excluding Japan) is predicted to grow a five-year CAGR of 69.8% to US$5.0 billion in 2021, according to the report by IDC.

But What Is AI?

While there are many ways to define AI, it is simply put: “anything written by a human that allows a machine to do human tasks.”

This is achieved with comprehensive programmes, data sets, and prescriptive algorithms which guide the machine to replicate human behaviour, actions, and decisions.

The contributions of AI may seem boundless, and anticipated opportunities plentiful, but there is much to be improved before it can be optimized to its fullest potential.

Amidst our venture in the use of AI to aid the digital transformation of businesses, there is still room to improve the technology.

The Human Touch

When it comes to AI, some people still jump to the myth of sophisticated human-like robots taking over.

However, what is often overlooked is that the interactive responses from these AI-run machines are simply pre-programmed algorithms mining through a vault of data sets, which then serves a pre-defined response.

While this works wonders for projection of data, other factors contributing to decision-making processes, such as personal experiences, business acumen, and creativity, are lacking. This prevents the wider application of AI into the creative service industries which operates beyond statistically sound outcomes.

For example, Chef Watson, the AI robot developed by IBM “created recipes using its flavour algorithm, with only limited human assistance in refining the flavour profiles”—the results were positive, but not without a dash of confusing recipes, like the Austrian Chocolate Burrito.

Quality of Data

AI is able to generate strong statistic-backed solutions based on the richness of the data vault it is equipped with.

However, this reliance of quality and hygiene of data left a crippling flaw—”the inherent bias that exists in user generated data.”

User-generated data is laden with personal opinions, cultural influence, and social background that may distort the semantics of the data. If the data is simply extracted without context, literal interpretation of it could lead to subpar performance.

Undeniably, use cases such as Chef Watson represent AI’s potential in supplementing human creativity and behaviours. However, for AI to operate independently without human direction, more needs to be done to help expand AI’s cognitive abilities beyond programming rules.

The Silver Lining

The good news is that markets, such as Singapore, have recognised these setbacks and are taking steps towards improving the use of AI.

To ensure that AI initiatives are market ready, the Singapore government has been partnering with industry experts and engaging in cross-sector partnerships.

Singapore has been rolling out trial projects to test out the operability of different smart city initiatives, from the smart lamps to smart cars, when subjected to different environmental stresses and situations.

Though the current tests are based more on environmental conditions instead of human facing scenarios, perpetual optimization and field testing should help with realising the use of AI in the future.

So What Now?

To date, AI has helped augment capabilities across many job functions and industries by reducing human errors.

It has also intelligently flagged bottlenecks and inefficiencies to be improved, highlighted opportunities by leveraging on empirical predictions, and filled in gaps in operations to boost the operating capacity of existing technology.

Looking to the future, greater optimization of existing AI capabilities is necessary for it to transition from a supporting technology to a competent replacement of human abilities, taking on some of the manual tasks to leave room for greater strategic thinking and innovation.


Leave a Reply