The Democratisation of Artificial Intelligence
Many of the most powerful AI tools are freely available to everyone
Artificial Intelligence will disrupt every industry; that seems to be pretty much accepted now. But rather than asking how, many seem to be waiting for some innovative company to invent the AI product that will solve their problem.
That is complacent. If someone invents the solution to your problem, they also invent the solution to your competitors’ problem, and you’re at no greater advantage. But of more concern is that the inventor of the next disruptive AI innovation in your industry might find they are so good at solving your problem, they become your competitor.
New AI tech is all very exciting. But if you’re interested in disruption – whether you’re hoping to be the next unicorn that blows apart a stale industry, or part of one of those industries and want to lead the drive for change – you need to look at how the tools of AI can be used to create that disruption. In other words, you need to be the one doing the disrupting.
AI for all
From a business point of view, probably the most important thing to happen in AI started in late 2015 when Google, Microsoft and Facebook, closely followed by IBM and Amazon amongst others, made their AI tools freely available.
We cannot overstate the importance of this. These hugely powerful tools, used, developed and backed by the world’s most advanced technology companies, became freely available to everyone. AI was democratised.
Since being made public, these tools have been greatly improved and received huge interest, and they will only get faster, smarter and more useable. AI is already helping engineering companies model new jet engine designs, oil companies predict where to drill for oil, drug companies indentify promising new areas for research. Big change is happening.
AI: The next big digital disruptor
Freely available AI represents a disruptive force that should now be familiar and potentially worrying to non-digital businesses.
This is all part of a trend of doing more and more digitally. Since the birth of the internet, disruption has followed a similar pattern: established companies have assumed that their size, infrastructure, legacy market position and supply chains were too big barriers of entry for startups to be a threat. The internet, then cloud, then data analytics proved a growing number of industries wrong. Each innovation gave digital competitors power to do many disruptive things with less infrastructure, lower risk and greater precision. AI is about to do the same for many of the businesses that have made it this far without ambitious digital transformation plans.
Amazon undercut retailers by cutting out physical stores and better targeting customer via data. Spotify eliminated the need for expensive data centres and vast IT teams by scaling rapidly through Amazon’s cloud. These companies could deploy new digital services rapidly, use data to target customers, experiment, fail fast and change tack at short notice. Industries such as retail and entertainment were often blindsided, slow to react and are struggling to catchup.
Build AI and take on the world. Or be taken over by someone else’s AI.
AI moves this forward in leaps and bounds. Companies already doing things digitally now have incredible power to derive AI based insights from all their available data. Equally, many companies yet to embrace digitalisation are about to find out about disruption the hard way.
Data analytics allowed companies to identify interesting patterns in data which could help them better target customers and understand operations – transforming online sales and marketing, and well understood production processes. AI takes this understanding, prediction and targeting possibilities into new fields, new marketplaces, with much more complexity and where much greater accuracy is needed – healthcare, insurance, energy, engineering.
Drilling for oil or building a jet engine still requires physical infrastructure, but that is no longer a total barrier to entry. Much engineering, design and R&D can now be done in the cloud using AI tools to model best outcomes, and many of the systems that underpin these approaches will soon be automated and managed remotely.
AI platforms will get exponentially faster and more sophisticated, meaning the requirements for someone to disrupt even complex industries are disappearing at pace. It will become very hard for anyone not using AI effectively to compete. Some clever startup with industry expertise and a few data scientists will turn old assumptions on their head sooner or later.
The tools of the digital disruptors
Google, Microsoft, Amazon and IBM are the main players offering a cloud based AI stack. Others can be easily combined. At the top level are AI services such as bots, image recognition, and natural language processing. Tools you can apply to your data to develop new insight, or integrate into new products and services.
Next is the AI infrastructure layer, which allows developers to build AI tools such as machine learning and neural nets using existing frameworks. Finally, there is the AI Framework layer which provides the fundamental building blocks of AI – open source software libraries such as Tensorflow – which can be used by data scientists to create bespoke AI as advanced as anything currently possible.
Whilst these are all important, it is those operating at this bottom layer, building truly bespoke tools from the ground up, that are likely to be the big disruptors of the next five years.
The end of the non-digital business?
The upshot of all this is that anyone can build AI tools in the cloud, which are as sophisticated as anything Google and Microsoft have ever produced. When we hear about IBM Watson diagnosing cancer – this isn’t some mystical service outsourced to some dark corner of IBM, it could have been done by anyone with the right data skills. My team could have done it.
But we shouldn’t be too hasty to condemn any business more than 15 years old to the recycle-bin of history. Retail and entertainment were blindsided by the sudden pace of change. Other non-digital industries have been warned, and many are taking the threat/opportunity of digitalisation seriously. Leaders in non-digital sectors, including companies such as BP, GSK, and Siemens, have excellent transformational digitalisation programmes in flight, with huge focus on the potential of AI in their business. Such companies have the confidence to make significant investments in digital, and detailed knowledge of their sectors. This should play hugely to their advantage, if they can drive the cultural change needed to thrive in this new world.
The technology barrier to entry is fast disappearing from all industries. Skills and expertise in using AI and related tools to derive insight from data is where the real competitive advantage lies. Companies which identify a problem that needs solving; understand the context, find the right data, apply the right intelligence and build the right solutions with the right tools will be the ones who bring about the next big disruption. Who hits on the big disruptive ideas first, remains to be seen. Either way, we are on the brink of the next data revolution.
Matt Jones is Lead Analytics Strategist at Tessella, Altran’s World Class Center for Analytics – using data science to accelerate evidence-based decision making, allowing businesses to improve profitability, reduce costs, streamline operations, avoid errors and out-innovate the competition.