The AI Race

The Rise of the Machines

What Are the Opportunities for Artificial Intelligence?

  • As a new factor in production, AI could double growth rates in developed countries by 2035.
  • Food producers could see costs cut in half with AI precision farming promoting a ‘green revolution’.
  • Extra income for vehicle owners and operators could be generated if autonomous cars are put into a shared fleet, or their computing capacity is used to mine for crypto currencies, when not in personal use.
  • Media consumption could rise as people in autonomous vehicles look for new forms of entertainment.
  • More land could become available in urban areas for competitive submarkets as the use of autonomous cars reduces the need for parking lots.
  • Insurance coverage could become commercial rather than individual and legal systems will have to adapt to revised liabilities as autonomous vehicles are adopted more quickly than expected.
  • Insurance companies could use AI and the Internet of Things to predict or mitigate against loss, switching their focus from loss adjusting to monitoring claims.
  • Creative artists and philosophers will be in demand as AI, Big Data and Machine Learning eliminate all but the most entrepreneurial, highly skilled and creative jobs in the economy.

What Are the Challenges for Artificial Intelligence?

  • The US$400 billion services industry could shrink as domestic robots and voice assistants take over the household chores.
  • Increased automation in manufacturing could contribute to higher unemployment as the need for labor reduces.
  • With the rise in use of Robo-taxis and other autonomous vehicles, consumers could stop buying their own cars and the jobs of truck drivers and public transit employees could be at risk.
  • Large consumer internet companies might have to rent personal data to use in their AI applications as people move towards fully owning their personal digital data.
  • If wireless technologies are proven to be carcinogenic, they could dramatically slow down the proliferation of wireless devices and the Internet of Things.
  • Governments could bring in regulations, or set up agencies to monitor companies that can’t regulate the interdependencies between larger computer networks.
  • Security and policy-making will need to keep up with the pace of AI changes to avoid breaches in trusted hyperscale sources.

At the heart of Artificial Intelligence is software powered by machine learning and deep learning (using neural networks) which can replicate, and will eventually surpass, human intelligence.

By 2025, Artificial Intelligence (AI) will be driving the biggest technological advancements across the auto, banking, compute, energy and health care industries.

Companies likely to harness the power of AI most successfully will be those with robust fundamental outlooks.

  • AI is forecast to achieve a compound annual growth rate of 28% from $692 billion in 2017 to $5.025 trillion in 2025 according to global research and advisory firm Gartner.

AI Business Value forecast 2017-2025

"The country that leads AI development will be ruler of the world."

- Vladimir Putin

From Sci-fi to AI

Scenarios that used to exist only in the pages of science fiction, are fast becoming a reality. With the increase of compute power, Big Data has developed via the proliferation of the Internet, mobile devices and sensors. Because the cost of storage has reduced, it’s now possible for deep learning algorithms to use all this data.

While AI is being used to solve fairly basic tasks, as the technology progresses over the next five to 10 years, it will become increasingly sophisticated. Each stage will have different implications not just for a range of industries, but for TCS (Tata Consulting Services) suggests that 84% of executives believe AI will be essential to competitiveness in the future. They also think growth rates in developed countries could double by 2035 as AI becomes a new factor in production.

AI applications are wide ranging, from autonomous cars and voice-enabled devices to bioengineering and industrial robotics. And the economic, social and political implications of AI could be profound.

Here we present the highlights from ‘The Artificial Intelligence (AI) Race’, the second chapter from Imagine 2025 –Themes, Opportunities & The Law of Accelerating Returns, an examination of the global drivers of parabolic change. To read the main Imagine 2025 Report in full, or for more information, Register Here.

There are three types of Artificial Intelligence: Narrow AI, General AI and Super AI.

01. Narrow AI

This is the most immediate application of AI as a single automated activity which outperforms human efficiency. Most applications to date are Narrow AI. We’re already using it for online purchasing and music recommendations, streaming services and high-frequency trading. With the help of IBM’s Watson, AI is also assisting doctors to make diagnoses. It’s worth noting that autonomous driving, a form of Narrow AI because it has a single functionality, is far more complex than other existing forms of Narrow AI.

Still in its infancy, AI technology will continue to progress. And many of the improvements will come through games such as Google’s (Deepmind) AlphaGo, based on the board game Go, and OpenAI’s Dota 2, a complex video game. These developments show how applying AI to other areas of society can improve efficiency and productivity.

At the Forefront

  • Google’s (Deepmind) AlphaGo Zero took 40 days to become the best Go player in the world entirely from self-play with no human intervention and using no historical data.
  • After OpenAI created a program to beat the world’s top Dota 2 players in 1-on-1 matches, the team’s next goal is to teach the system to play 5-on-5 matches.

40 days: AlphaGo Zero surpasses all other versions of AlphaGo and, arguably, becomes the best Go player in the world. It does this entirely from self-play, with no human intervention and using no historical data.

02. General AI

Artificial intelligence that can perform any intellectual task a human can is called General AI, or sometimes human-level AI. It’s harder to achieve than Narrow AI and is not focused on a specific activity. For AI to be as intelligent as a human, technological breakthroughs in both hardware and software will be needed to replicate the human brain’s computing power. Once the hardware is there, the software must develop. There’s already been strong success with deep neural networks in AI, for example AlphaGo Zero.

03. Super AI

Super-intelligence, also called the Singularity, is achieved when AI becomes much smarter than humans. Again, significant improvements in hardware and software will be needed to advance Super AI. But we believe it’s important to look at this through a future lens as technology can change exponentially compared with humans who generally think in a linear fashion.

And intelligence, as we understand it, has been increasingly non-linear. This suggests that AI is just the next stage of evolution on earth. Brain-machine interface (BMI) is not new but the research so far has focused on using it to restore functions to people with disabilities. For instance a cochlear implant to restore hearing is already in use. As more funds go into the race to hack the human brain, the idea of integrating AI into ourselves could become more than just science fiction.

At the Forefront

  • Elon Musk’s Neuralink start-up is looking to create a brain-to machine interface. It’s developing neural lace, an ultra-thin mesh implanted in the skull, to increase the storage and processing power of the human brain.

Get the full report today.

For additional insight into the forces of change impacting the future, read the full Imagine 2025 report.