In the race to embrace artificial intelligence swiftly, companies are investing millions of dollars to stay competitive. In most cases, AI development is leading business strategy — not just assessing the business needs but building new capabilities to accelerate strategic AI transformation. However, experts believe that it should be the other way round — companies should let their business strategy lead AI deployment. But despite all the hype AI is still in its infancy stage, an AT Kearney report reveals.
Some 5,000 AI-based startups which have been launched since 2014 have attracted a total of $40 billion from venture capital firms and the investment arms of leading technology companies. Capital is flowing across the sector, with autonomous vehicle systems landing the most money, followed by RPA technologies.
In the race to embrace AI fast enough, enterprises have to grapple with the hard truth — slower ROI. As the use cases get narrower and technology penetration deepens, demand for AI visionaries and computer scientists is shooting up. Technology experts are now beginning to harp about the transformative power of emergent technologies — consumer-facing AI applications, cloud computing and the internet of things. But despite all this, a report in Foreign Policy emphasises that the final payoff is a decade away.
The reason seems to be that general-purpose technologies can be used in so many ways, but their adoption takes a long time to reach critical mass.
Is AI A General-Purpose Technology?
General-Purpose Technologies (GPT) are known as the engines of growth. For example, the steam engine and electricity played that role in the past in enhancing productivity and advancing economies. Similarly, semiconductors and computers are playing a similar role today, bringing substantive economic gains which are reflected in productivity statistics too. GPTs are known for their technological milestones and overhauling the economic landscape, thereby creating business value.
However, as Erik Brynjolfsson, and Daniel Rock put it in their NBER Working Paper, one of the key subsets of AI — machine learning has not seen the full enterprise adoption. A key point presented by Rock and Brynjolfsson full effects for general-purpose technologies will not be realised until complementary innovations are developed and implemented. The paper further reveals that adjustment costs, organisational changes and new skills have to be factored in for a successful AI to drive economic growth. It also drives the market value of firms and national statistics can fail to measure the full benefits of the new technologies. In other words, for AI to reap the economic benefit, it will take time for the true potential of a GPT, in this case, AI to become clear. Thereafter, it takes more time for firms to decide how to adapt production processes accordingly.
AI Has Shown Promising Signs
AI technology is mired with legal and regulatory concerns that range from data security to anti-trust and liability issues. This is slowing the development of deployment of consumer AI applications. Currently, despite the hype around autonomous vehicles, automakers and autonomous tech providers are dealing with delayed deployment. The paper emphasises even though there is an increased adoption of certain consumer-facing AI solutions, there’s absolutely no knowing whether consumers will be ready to embrace the full breadth of AI technologies which will be introduced to the market.
AI — The Most Important GPT Of The Era
Technological optimism is abound when it comes to AI and investors. Researchers are also putting in exponential amount of money and resources to increases its use cases and drive widespread adoption. But there’s a negative side to all this — its benefits are being reaped only by a small section of the economy.
Here are some of the reasons outlined for AI as GPT:
- Adoption to drive innovation across sectors
- Will generate major social benefits and improve welfare and productivity
- Spillovers throughout the economy as was with previous general-purpose technologies
Try deep learning using MATLAB