What Contributes to AI?

“Any sufficiently advanced technology is indistinguishable from magic.”  – Arthur Clarke.

When we think of artificial intelligence (AI), we think of machines, computer systems, algorithms, and models trying to get closer to human intelligence. Today, this space straddles many areas, like natural speech recognition, expert systems, machine learning, general AI applications, and more. The future, however, is mapped towards singularity:

  1. There would come a time when AI would exceed our expectations and assumptions.
  2. We would meet the realm of Strong AI.
  3. Humans could leverage this form of intelligence at an unprecedented level and in exponential ways.

But to get there, a lot of pieces of the AI puzzle would have to come together. Here are a few important ones.

Artificial Intelligence (AI) and Big Data

AI is nothing but intelligence that learns, re-learns, and reinforces its learning through the data that it gets. This data can either be specifically fed into the modes or designed to collect this data independently. Hence, the volume and quality of data are extremely important in determining the efficacy and relevance of AI for any purpose. 

Big data is a new form of information asset, and it requires massive processing models and the right computing muscle to take care of its flow, velocity, scale, and density. For Big Data to expand and extend human intelligence through AI, we would need to master many dimensions of Big Data. They would include, but would not be limited to, speech recognition, IoT and sensor data, image recognition, unstructured data, and real-time data of many formats. 

The space of Big Data is growing rapidly and in multiple directions. As per reports from MarketsandMarkets, it can grow from $138.9 billion in 2020 to $229.4 billion by 2025 and enable companies to increase operational efficiencies and reduce costs. Of course, that should help the growth of AI too.

Artificial Intelligence (AI) and innovation

AI is much more than the fundamentals of the technology that power it. Its real impact and possibility lie in the way it is augmented consistently. It has to get the right research and creative impetus to explore new models, processing modes, algorithm designs, and enablers. There are many constraints for AI, as we see them today. They deter many enterprises from leveraging AI with its full potential. Gartner’s 2019 CIO Agenda survey shows that only 14 percent of global CIOs have already deployed AI. It signified that by 2024, 50 percent of AI investments would be quantified and linked to specific key performance indicators to measure return on investment. About 13 percent struggled with governance issues, and 42 percent feared the unknown in understanding AI uses and benefits. Then there are barriers around vendors with complexity, integration hassles, and capabilities. About 20 percent had privacy concerns, and six percent wondered about the risk of liabilities. Innovation has to cross these barriers and catapult AI to a plug-and-play level for enterprises. Another barrier to AI adoption is quality data (18 percent), as pointed out by the O’Reilly annual AI Adoption in the Enterprise 2021 survey. These barriers will have to be surmounted with the right innovation in academic spheres and the right governance models.

Artificial Intelligence (AI) and talent

The first reflex thought that most professionals feel when they hear of AI is displacement because of AI. This is not just because AI can take over repetitive jobs. But because AI applications would need a new set of skills and mindset, professionals would have to get ready for it. According to a Gartner Research Circle survey, about 56 percent of respondents felt that acquiring new skills will be required to do both existing and newly created jobs. Interestingly, the O’Reilly annual AI Adoption in the Enterprise 2021 survey says that demand for AI expertise has exceeded supply. Companies are feeling the skills shortage most acutely in the areas of ML modeling and data science (52 percent), understanding business use cases (49 percent), and data engineering (42 percent). Another reflection of the AI skills gap is the finding that the percentage of companies with AI products in production over the last year (25 percent) is flat when compared with 2020 (26 percent) and 2019 (27 percent).

Artificial Intelligence (AI) and wider applications

AI is pointless if the intelligence derived here is not fit for the right purpose. About 42 percent of respondents aren’t fully understanding AI benefits and use in the workplace, as per Gartner’s reckoning. Quantifying the benefits of AI projects is another problem for IT and business leaders. When 26 percent of CIOs have a problem finding the use-cases and 25 percent are wrestling with defining the strategy, then a lot remains to be done on the usage aspect of AI. 

According to the McKinsey ‘The State of AI in 2020’ report, 50 percent of companies have adopted AI in at least one business function. AI adoption is highest within the product or service development and service-operation functions. In terms of dominant AI use-cases, manufacturing shows 15 percent with yield, energy, and/or throughput optimization and 12 percent with predictive maintenance. In HR, the use is more on optimizing talent management (10 percent) and talent management (7 percent). Marketing and sales are inclined towards customer-service analytics (17 percent) and customer segmentation (14 percent). Strategy and corporate finance areas find their usage in capital allocation (8 percent) and mergers and acquisitions support (6 percent). Logistics-network optimization tops supply chain management with nine percent, and so does inventory and parts optimization (9 percent again). 

We have barely scratched the tip of the iceberg here. To fully leverage the power of AI, we need to explore many more applications, but only when we have models and tools that give that kind of confidence.

Artificial Intelligence (AI) and investments

A lot hinges on how much the top leaders of an organization feel excited about AI. Investments will drive more innovation and more capabilities ahead. For instance, the McKinsey report shows that AI top performers invest more of their digital budgets in AI than their counterparts and are more likely to increase their AI investments in the next three years. As to the distinct aspects of these high performers, there is another point worth noting. The ability to develop AI solutions in-house, instead of buying solutions – and they typically employ more AI-related talent, such as data engineers, data architects, and translators, than their counterparts. These companies have also built a standardized end-to-end platform for AI-related data science, data engineering, and application development. So enterprises would have to find that kind of enthusiasm as budgets to leverage AI strongly. 

Artificial Intelligence (AI) and processing power

AI depends a lot on how models use data, learn, and apply their intelligence. That is where AI would need higher and wider processing alternatives. Especially in the area of real-time data. The McKinsey report also talks about the adoption of deep learning—a type of machine learning that uses neural networks and can sometimes deliver superior results. As of now, just 16 percent of companies seem to have taken deep learning beyond the piloting stage. Only high-tech and telecom companies are leading the charge here, with 30 percent embedded deep-learning capabilities. Opening up to new technology paradigms would be a precursor to real growth in AI.

Artificial Intelligence (AI) and improved algorithms

Companies are still not sure about handling a lot of AI ramifications. That’s where the algorithms need to evolve and iron out these kinks. In the McKinsey report, we see how companies are still worried about abort risks like explainability (41 percent), privacy (39 percent), equity and fairness (24 percent), and cybersecurity (62 percent). In fact, as per the O’Reilly survey, mature organizations checked for unexpected outcomes or predictions, interpretability and transparency, and model degradation. Privacy and fairness, bias, and ethics were also ranked above 50 percent in terms of concerns. We need algorithms that can address these issues and bolster the excitement that enterprises have towards AI. If all these ducks fall in a row, AI would swim to new shores of possibility and impact soon.