Speeding towards a better world with Exponential AI

People have been getting worked up about artificial intelligence lately because suddenly, it appears, AI and the algorithms behind it have become a lot more capable. In fact, they have been getting exponentially better for many decades. This advancement is no different to your milk “suddenly” going off, which occurs because a few bacteria find their way into the carton and divide in two. Each cell division takes about 20 minutes. Toward the end of the spoilage, the milk is ¼ bad, then ½ bad, then all bad – giving the impression it suddenly went bad in that final 1 hour. It’s the same with AI developments magically appearing more capable, but this is just a result of early growth not being as noticeable.

Although research into AI began in the 1950s, it wasn’t until 1997 that a computer beat the world grandmaster at the game of chess. Then, in 2011, a computer system beat the world’s best human opponents at the game of Jeopardy – which involved the use of “natural language processing.” In 2017, a new type of computer system beat the world No. 1 ranked human player in the game of Go. In March 2018, Microsoft announced[1] another language-related milestone: human parity in English-Chinese translation. In these examples, we observe an increasing cadence of events where AI skills match or surpass human skills: 47 years, 14 years, 6 years, 1 year and on and on.   

Businesses will undoubtedly face challenges and opportunities as automation exponentially improves on the current human production of goods and provision of services. In some cases, computers will be faster, better, cheaper and safer. McKinsey[2] estimates a potential for AI “to create between $3.5 trillion and $5.8 trillion in value annually”.

Three types of technological development have combined to produce systems which exhibit ever-increasing skills: processors, networks and algorithms. The exponential increase in processing power has been named Moore’s Law after Gordon Moore, one of the founders of the chip maker Intel, who predicted an annual doubling of the number of transistors on a computer chip – starting in 1965. Less well known is Reed’s Law, which points out the exponential increase in the value of the connections in a network (of people) due to the potential for forming groups (particularly for collaboration). For example, connecting two people makes a pair, but four people can form 11 different groups (including 6 pairs) and eight people can form 247 groups (28 pairs).

The rise of the internet and the rise of social media on top of it has greatly accelerated our collective ability to work together to advance technology. Finally, Kurzweil’s Law of Accelerating Returns maintains that increases in processing power and networking fuel more rapid development of designs for even more powerful processors and efficient networks via better algorithms – so we can go faster.

From driverless cars to robotic workers, the AI future is already here and getting smarter - fast. Technological developments are radically changing the future of life and work. We see a lot of convergence, multi-tasking and free access to information and knowledge. AI will open our abilities to solving a new scale of problems and provide solutions for the betterment of humanity.  

Businesses now have opportunities to use more computing power, wider networks and better algorithms to create customer value. These developments can be seen in numerous examples of disruptive technology, such as;

  • Blockchain – an immutable ledger based on cryptography (Bitcoin, Ethereum)
  • Driverless Cars and Pilotless Air Taxis – image recognition, robotics (Uber, Zephyr Airworks)
  • Recommendation Engines – collaborative filtering (Amazon, Netflix)
  • Voice User Interfaces – natural language processing (Siri, Cortana, Alexa, Google Home)

It is easy to comprehend the value that AI can bring to organisations, but it’s not always clear where to start. The key to efficient Machine Learning and AI is data, which allows algorithms to “learn” and become more intelligent over time. Therefore, a manageable starting point for all business is to become fully data-driven, whether this is using a Marketing Automation platform, intelligent CRMs or AI for competitive intelligence. Resource restrained organisations have the option of partnering with experts instead of hiring in-house specialists, to begin the journey to become data-driven.

If, or perhaps when, these technologies become mainstream, it may be possible to simply, verbally “ask” the ubiquitous intelligent infrastructure enabled by AI to conduct transactions, transport us to our desired destinations and make helpful suggestions.

Including, presumably, “Don’t forget to buy milk.”  

 

[1] https://blogs.microsoft.com/ai/machine-translation-news-test-set-human-parity/

[2] https://www.mckinsey.com/global-themes/artificial-intelligence/notes-from-the-ai-frontier-applications-and-value-of-deep-learning?cid=soc-web#part3

Subscribe to the blog