While there have been talks of AI posing a grave threat to humanity, I believe we will always have enough control and foresight through human intelligence to manage and maintain the perceived risk of AI going out of control. Of course, AI could be used for destructive purposes in the realms of physical and digital security, but again we need to counter these issues using the technology itself.
From an employment perspective, there is surely going to be a fairly rapid change in how AI and automation will impact the work environment, especially at the lower end of the skill spectrum.
We are already seeing this impact across various industries and roles, such as the retail cashier, accounting, customer service operations, etc., where AI can reduce the dependency on human intervention in a process flow while increasing productivity and efficiency.
This is akin to how automation had reduced the need for manual labor during the period of rapid industrialization. It is also essential to understand the evolution from automation to intelligence, where automation had systemized repetitive tasks, and AI has taken the next quantum leap to make decisions based on statistical modeling to various exceptions or situations that come up during a repetitive process or task.
As a result of the combination, we are now seeing large production facilities or customer service operations being almost unmanned, and we can extrapolate this to seeing a day when we can have entire organizations operating through AI with insufficient human involvement.
So, whether its chatbots or IoT, automation and AI will disrupt the job market from a skill standpoint, and this will result in education systems and skill development moving upwards regarding the skills that people need to bring to the workplace.
However, on the other hand, I think we will see new age industries come up as well. The e-commerce industry has disrupted the traditional retail sector and continues to grow at a rapid pace, bringing substantial investments and thousands of jobs in several countries.
A similar scenario is also seen in the case of aggregators, whether in food, travel or anything else. Even the space industry or the energy industry, have been able to make significant shifts in their commercial and operating models due to the adoption of technology and AI.
We should also realize that all the productivity gain from technology and AI will free human and financial capital, which will help us spawn innovation and new industries, which will drive economic and job growth.
The time is right; the time is now: Businesses NEED to get AI ready
Businesses need to look at AI from a Tier perspective. Irrespective of the industry, you can break all companies up into three tiers – Enterprise, the Mid Space and the Small and Medium Businesses.
Typically, larger enterprise organizations tend to have a strong history of data management and an established platform of systems, which is a prerequisite for anyone trying to utilize machine learning or artificial intelligence in any sense. If you look at enterprise companies, they usually use Tier 1 ERP’s like SAP, Oracle or Microsoft Dynamics for data accessibility and storage in conjunction with other systems like CRM, HR, BPM and other platforms. The data in all of these systems become the source data for any machine learning or artificial intelligence exercise.
The mid-segment companies, on the other hand, tend to be in a state of evolution regarding systems. These companies often start with analytics of existing data rather than deploying AI, though there is always an interest in AI as part of their growth strategy. However, it is crucial to note analytics is also evolving to include AI regarding mining insights and suggesting actions, beyond traditional dashboards and reports.
And, at the bottom of the spectrum, small businesses are adopting cloud technology from vendors, such as Microsoft, which has inherent public AI capabilities. Also, several small companies are building their business models around AI capabilities, and therefore they will scale up on technology use within their businesses, irrespective of their size or ability.
Another lateral view of this situation is that particular industries have a greater need for AI and machine learning and a more prominent ‘use case’ for it. For example, a retailer with a more extensive database of transactions has a greater need to use artificial intelligence to understand various aspects of their customer base and operations better. AI could help in predicting consumer purchase trends, store inventory replenishment models, merchandising inputs and so forth. Therefore, specific industries have a higher propensity towards usage of AI or ML as compared to others.
We can broadly see that the larger the volume of data and transactions, the higher the likelihood of the company utilizing the information to figure out trends and patterns to unlock customer demand and find bottlenecks in their business, whereas, with lower volumes of data, it is more difficult to predict patterns and build logic around data sets.
In the context of businesses being ready for AI
I feel AI is getting simpler as well in terms of programming and data science itself. Even though it is complicated for the layman, it is becoming more accessible and understandable. It is gradually progressing towards being more interface-driven, and that is eventually going to lead towards the adoption of AI across the board.
The future of search giant Google is AI, says their CEO Sunder Pichai. In fact, a conceptual shift from ‘mobile first’ to ‘artificial intelligence first’ is already evident as they lay greater emphasis on machine learning and voice recognition in 2018. While at its current stage AI is more of an enterprise and mid-enterprise play, it is going to seep into the bottom layer reasonably quickly.
As I see it, 2018 will be the year when the AI debate will transition from its potential to its utility. While there is no definitive ‘doom or boom’ prediction about AI’s future, I believe that AI will be the enabler of the next stage of human evolution.
This article was published in AMEInfo