Artificial intelligence is a topic that’s been discussed for decades, but it’s an industry still very much in its infancy - we’re only seeing the beginning of its capabilities. There are areas where AI has become heavily relied upon - such as algorithmic trading - but, in general, the broad adoption of the technology is still marginal. As an industry, it’s a toddler you could say, but we’re at a point in time where we can expect to see it grow up - and fast.
There have been notable achievements and breakthroughs throughout the years which we can look at to get a better understanding of where AI is at today. First came expert systems that were adopted in the 70s and 80s for use in our cars, PCs, and other forms of manufacturing, but which failed dramatically when applied to fields such as healthcare, so hit a barrier in terms of their exponential adoption.
Google’s search engine and Amazon’s recommendation system were the masterpieces of the next AI wave in the 90s which introduced today’s ‘pattern recognition’ boom. This is all about AI learning to recognise features and patterns in complex data even where humans fail to identify them. The biggest success stories here are in:
- Natural language processing, from speech recognition and machine translation, to automated text summarisation and question answering.
- Image, video, and other signal processing for detecting complex patterns, such as activity trackers learning to detect different types of activity from motion signals.
An important milestone was the highly publicised machine over man triumph in 1997 when IBM’s Deep Blue chess computer won over the reigning world chess champion Garry Kasparov. This was symbolically significant because it was one of the first demonstrable examples of a machine outperforming a world leader in its field. The Deep Blue victory established an understanding that AI could be used to solve very complex problems. If it could beat the best chess player in the world, what could it do next?
Since then, AI has found its way further into online user experiences and optimisation of online ads, but hasn’t been adapted as fast as many may have predicted 20 years ago. However, recent advances in deep learning, exemplified by Google's AlphaGo surprise win over the world’s elite Go players in 2016, signal that a new generation of AI algorithms are making their way into the market. This suggests that the next 20 years will see an acceleration of the importance of AI almost any industries.
We’re now seeing three pillars of AI markets which are all developing in different ways, with various companies operating within them.
- The ‘AI as infrastructure’ market where companies directly sell AI services or platforms.
- The ‘AI as a vertical’ market where highly-skilled teams apply AI to a specific vertical, directly competing with established players, e.g. in finance, life science, energy etc.
- The ‘AI as a growth driver’ market. Here, AI technologies are applied - often in a very bespoke way - to specific problems in a wider industry, e.g. object recognition for detecting hurdles in a traffic lane.
If we look at the forecast of AI across the next five years, the biggest trend impacting the corporate world is the importance of external data and how AI will be used to incorporate this into more proactive decision making processes. External data is one of the biggest blind spots in corporate decision making today, with many executives making decisions primarily based upon internal insights. This is a very reactive approach because internal data is a lagging performance indicator. It is looking at the result of historic events that took place in the past - weeks, months, quarters, sometimes years in the past.
In external data, however, you can find many forward-looking insights about your entire competitive landscape. By monitoring job postings you can track - in real-time - the appetite for investments among competitors, partners, distributors, and suppliers. You can also harness insights into how competitors spend their online marketing dollars; do they increase their spend in Europe or are they doubling down in North America? By mining social media, you can pick up on changing trends in consumer preferences informing investment decisions in existing or new product lines. By analysing external data, executives can find forward-looking insights and indicators to help them stay on top of changes in their competitive landscape and to be proactive in their decision making. We call this approach OI (Outside Insight) and over time we believe the need to analyse external data will grow into an entirely new software category analogous to what BI (Business Intelligence) is to internal data.
In saying this, the ultimate potential market for AI is very large and will extend far beyond its current scope. The industries AI is having an impact on will continue to expand with transportation, food and drink, healthcare, finance and risk assessment likely to be the most transformed by new approaches. We’ll also continue to see even more successful targeting outside of Adtech; specifically moving into politics (Cambridge Analytica, Palantir), journalism (Buzzfeed, targeted content farming) and healthcare.
Three key factors driving AI going forward are:
- Exponential growth of cloud-based computing power (per dollar)
- Further advancements in AI techniques such as deep learning
- Continued growth of available data (internal company data, external online data, data generated by IoT)
Combined, these three factors will make AI stronger, more reliable and more relevant for an increasing number of decision makers across functions in any or industry. As such, AI will play a meaningful role in the total corporate IT spend within the next decade. It’s reasonable to expect it to grow into the hundreds of billions of dollars, if not more.
Although the development of AI is at an exciting stage, there are still challenges related to the experience and skill required to design new systems. There will be big changes in the near future, but the best is yet to come.
Jorn Lyseggen, Founder & CEO, Meltwater
Image Credit: John Williams RUS / Shutterstock