By Prabhat Manocha, Associate Director, Government – Sector, IBM India / South Asia
The advent of digital technologies, social media and Internet of Things is resulting in huge amounts of data being generated. It is a famous saying, Data Never Sleeps. Some estimates and market reports suggest that by 2020; 5,200 GB of data would be generated every year for each individual on the planet. And, this data exists in various forms, shapes and sizes. For e.g. every minute, over 2 million photographs are shared, 4 million YouTube videos are viewed, over 12 million texts and 150 million emails are sent. Phew!
The irony, however, is that despite this information overload many CIO’s still take decisions based on insufficient data. You might ask why?
While there’s no shortage of data out there, the ability to access and act on it whenever, wherever, and however is limited.
This is where AI can help!
With AI, we can:
• Understand: Any form of data, either in the form of video, sound, blogs, image and any other.
• Reason: Provide context to the data
• Learn: Continuous learning and updating through machine learning capabilities
• Interact: Interact with humans using Natural Language Processing. This can break the barrier between human and machines. Infact, it is estimated that by 2021, at least 20% of citizen conversations would be done using AI chatbots.
Not surprisingly, AI is gaining popularity, as evident from worldwide spending on Cognitive and Artificial Intelligence Systems which is forecast to reach $77.6 billion by 2022, nearly 3x times its 2018 forecast.
Taking a closer look at the AI efforts, Governments are focusing on sectors that are envisioned to benefit the most from AI in solving societal needs. Some of these are:
• Healthcare: Increased access and affordability of quality healthcare,
• Agriculture: Increase in farmers’ income, farm productivity and reduction of wastage
• Education: Improved access and quality of education
• Smart Cities and Infrastructure: Efficiency and connectivity for the burgeoning urban population, and
• Smart Mobility and Transportation: Smarter and safer modes of transportation and better traffic and congestion problems.
Let’s look at some of the use cases across some of these sectors:
Agriculture: We have worked to build precision agriculture application for various districts of India. With the help of crop data base, satellite images and weather information we can predict and update: Ground water deposits, crop diseases, natural calamities well in advance to take preventive actions.
Health: AI is being used in Life Sciences where it has the potential to reduce the time a new drug take to reach the market. A recent study found a correlation between burn injuries and cancer; with the use of AI, one is enabled in reviewing millions of pages of scientific literature in less than two months and single out mechanisms for further study. With this we should able to reason and find the root cause analysis that would help save many lives.
Education: As students globally struggle to complete their courses on time and Universities shorthanded at building suitable curriculum, AI is being used to make learning more personalized. Additionally, as teachers struggle to balance busy work and course schedules with the demands of advanced learning, AI-based tools can help teachers save time through effective planning.
Challenges and Concerns
However, there are some challenges to adopting AI that is effecting its pace and extent. As per a McKinsey study, only 21% say their organizations have embedded AI in several parts of the business, and only 3% of large firms have integrated AI across their full enterprise workflows and the reason for this is:
• Shortage of Skills: Globally skills necessary to tackle serious Artificial Intelligence research is not easily available.This requires investments in AI-relevant human capital and infrastructure to broaden the talent base capable of creating and executing AI solutions to keep pace with global AI leaders.
• Bias: AI systems are only as good as the data we put into them, and bad data used to train AI can contain implicit racial, gender, or ideological biases, as a result, biases creep their way into AI systems. More than 180 human biases have been defined and classified, and any one of themcan affect how we make decisions. Bias in AI systems could erode trust between humans and machines that learn. AI systems that couldtackle bias would be the most ideal and successful.
• Services available on the Cloud: Technologist and researchers have been building innovative application and services to be available “as a service” to be consumed via API; many government departments still have apprehensions about consuming these AI models from the Cloud.
• Policy: Despite the obvious opportunities for efficiency and effectiveness, the role of AI government policy and service delivery remains contentious. For example, when is it acceptable to use deep-learning models, where the logic used for decisions cannot possibly be explained or understood very easily? Citizens generally feel positive about government use of AI, but the level of support varies widely by use case, and many remain hesitant.
AI is an area where scepticism is high, its adaptation, methodology and investments might be argued, but smartly targeted AI and machine learning tools, with well-deployed algorithms fuelled by huge data sets, can drive lasting improvements across various social delivery applications of governments.
As per the survey done by McKinsey suggests, AI adoption could raise global GDP by as much as $13 trillion – which is 1.2% additional GDP growth per year – by 2030… Now isn’t that a dream made reality?