With the recent dispute between ‘tech-gods’ – Elon Musk and Mark Zukerberg regarding AI brewing, we decided to tackle the topic of AI and the impact it may have on your business and life.
Artificial intelligence (AI) is rapidly becoming the biggest issue on the agenda for many businesses. The current speed of development, the sheer range of possible applications, and the potential impact of AI suggest that it’s time for us to pay attention. So, what are the questions we all should be asking?
What was before AI?
Before we think about AI it is useful to look at its logical precursor that many organisations and people are still coming to terms with. There has been a tech trend for the past several years called Big Data. Big Data has been an attempt to draw all of the disparate data sources we have together and then to try analyse it using detailed formulas to find trends or patterns in the data. Then we have attempted to understand what these patterns may mean.
In reality Big Data has ended up being an exercise in frustration for most sites that have attempted it – because the sources of the data were so diverse. It was an exercise in patience, frustration, and BIG project budgets to pull the data together before it could even be analysed. The few who managed to get this right then spent lots of energy designing the formulas that would look for the patterns and trends in the data. The final few who managed to get some useful trend data then hit the hurdle of wondering how to adjust business’ strategy to take advantage of the patterns identified. Big Data has been tough to do well.
Labels applied to AI
Artificial Intelligence is picking up the gauntlet and moving on from what Big Data tried to achieve. The labels and names being applied to the field are not generally “Artificial Intelligence” instead we are talking about “Deep Learning” or “Machine Learning”.
These labels don’t address the expression and outputs of the Intelligence process but rather redirect our view toward how this is being done. This is useful because it helps us understand that true AI is the end result of a technology development process that gets the best out of the way machines learn. Our focus on how machines learn will enable us to mitigate against the doomsday scenarios of science fiction. Noted people like Stephen Hawking and Elon Musk raised their concerns in 2015 about the way in which un-tempered growth of Artificial Intelligence could undermine the development of our society. This seemed to come to fruition in Korea in early 2016.
An AI developed by Google called AlphaGo beat the world Go champion (a game it was believed a computer would never be able to master) in a round of 5 games held in South Korea. The significance of the victory was in the way in which the AlphaGo applied its AI. It used a process called Reinforcement Learning. Simply this means that AlphaGo took every possible decision it could take and even though it actually only executed one option it processed the learning of each possible alternative, thereby becoming exponentially more intelligent with each move.
Eventually, this process allowed AlphaGo to execute a move, in move 37 of the second game, that had never been played before. The significance of the move was that it was the closest we have ever seen an AI come to executing what we generally call human intuition.
Facebook has taken a different approach to AI. It is focusing on developing an AI that will teach other AI. This process will accelerate the way in which Facebook’s AI becomes progressively more intelligent and effective. Facebook uses its Artificial Intelligence engine to identify faces in pictures, drive the selection of what content from your network is pushed onto your timeline, selecting more relevant targeted advertising, and other actions that will make your activity on the network more effective and ultimately generate more revenue for the business.
The incredible thing about Facebook’s AI is that it will be open sourced and made freely available to anyone who wants to use it.
Artificial intelligence Open Sourced
Facebook is following an industry trend in open sourcing its AI engine. Google has a machine learning library called TensorFlow that it open sourced on 9 November 2015. TensorFlow was designed to drive Google’s image search ability – especially the ability to run a search using an image and not text.
In December 2015 Elon Musk and Sam Altman (a leading Silicon Valley Venture Capitalist) created OpenAI, an initiative that head hunted the top AI experts in the world from Google, Facebook, and other companies. Their objective is to create an open source AI development community that replicates the process that gave rise to the community that built Linux. This think tank will tackle the biggest challenges in AI, and as they solve them they will make them freely available to the rest of the world.
Artificial Intelligence has stepped off the pages of science fiction novels and is developing at a breakneck speed in our world today, it is not an element of some utopian / dystopian future.
But, what does all of this mean for a business person today, in our current world of work?
Below we tackle 10 key questions for anyone who in new to AI and its possible implication and application in the modern office.
1. Whats the fuss about?
AI will change the philosophy, practice and management of business. It is beginning to transform businesses and replace even senior management and leadership roles. CEOs must invest the necessary time and attention to understand what AI is and where the opportunities are
2. What’s its potential?
Start by educating yourself about its potential and undertake an internal analysis of where it could be deployed and what competitors are doing. Take a broader perspective of the potential roles AI could play from smarter production management, to customer targeting and broad-based decision making.
3. How fast is it moving?
The pace of AI development has caught most unaware. Large scale investment is being made by companies like Google, IBM, Microsoft, Uber and Baidu. In some cases, firms are literally “betting the ranch” on AI. We might soon see anything from robot lawyers to ultra-intelligent mobile personal assistants.
4. How deep should we take it?
Many firms are looking at relatively narrow deployments to automate rule-based decision making and predict future demand and customer behaviour from accumulated data. Others are looking at much broader uses such as intelligent HR, finance and legal advisors and real-time data analytics of live transactions.
5. Could it take the CEO’s Job?
Perhaps some of the most extreme applications include the creation of “human free” automated businesses where everything from strategy to operational processes are embedded “in the system”. These so-called “distributed autonomous organisations” could become increasingly common. They are already being used to manage millions of transactions – e.g. the automated dispute resolution systems in online auction platforms.
6. Who should lead?
This is not just another IT project. Artificial Intelligence has a much broader role and potential impact – so it should be the responsibility of the CEO, COO or business transformation head to drive the identification, piloting and application of AI solutions across all aspects of the business.
7. What would success look like?
The mantra should be to “fail fast and cheap” – bringing in suppliers, customers and other value chain partners early on to see if there is commercial merit in an idea. There can be as much learning from a failed project as a successful one.
8. How do we preserve the feminine?
Organisations and individuals display a mix of feminine and masculine characteristics. The challenge is to maintain the feminine in the human face of the organisation, and avoid the tendency for AI systems to display a more masculine and “robotic” persona.
9. How will staff respond?
AI is already being deployed to automate clerical, manual and semi-skilled labour and is now supporting and replacing professionals in domains such as engineering, medicine, legal, and accountancy. What’s your strategy for overcoming resistance and helping those who will be displaced by AI? How will you ensure those making it to senior positions have the necessary experience and expertise if staff numbers further down the pyramid are being reduced?
10. How do we address social impact?
There is a growing concern that unemployment levels could rise on a permanent basis. Research estimates suggest AI could replace 30-80% of all current jobs in the next 5-20 years. So, should we support the ideas of a guaranteed basic income for all – as is being explored by Finland, Canada and the Netherlands?
The pace of development and potential of AI mean that it is not something we can afford to put off – it is perhaps the single most important area of decision making business leaders will face over the next few years. The depth of a company’s understanding of AI could be the crucial differentiator between success and failure of firms in a fast-changing world.