This year, AI will be everywhere, whether you like it or not. From next level recommendation marketing and virtual agents to self-learning data analytics, there’s no escaping the touch of the machine in our everyday lives. AI is going from a next gen technology used by Google and Apple to something we encounter regularly – from chatbots to advanced analytics voice-activated assistants.
If you’re feeling a little uncomfortable about the idea, take solace from the past. During the original industrial revolution, not everyone bought into the idea of steam power. A group known as the Luddites went around smashing textile machinery because they were afraid they would be replaced. Of course, as industrialisation took hold and improved society as a whole, the Luddites became synonymous with backwards thinking.
We’re still in the early phases of our own industrial revolution – many of us understandably fear that our jobs are at risk, especially when we read that half our jobs may go to AI. But we’re already seeing the same patterns take place as the first industrial revolution, where AI isn’t replacing us at all, but instead allowing us to tap into our more human strengths.
Creativity, social intelligence, innovative thinking – these are just some of the skills technology can’t replicate. What it can do is give us more time and space to use them, by taking over repetitive tasks and sifting through large amounts of information. Just look at how AI is being rolled out in the medical world, not replacing doctors but giving them the power to make faster and more accurate diagnoses, help more patients, and offer more personal care.
The future of AI is one of coexistence and augmentation, where robots work with us to make us better at what we do.
If you’re looking for a yardstick on how much of a foothold AI is getting in South Africa, you need look no further than Michael Jordaan staking his retirement on NMRQL Research, an investment management company powered by machine learning. Compared to Bitcoin, which seems perennially stuck in the hype phase, AI is attracting credible investment from major players.
Despite this, there are too many people and organisations who don’t have the first clue of how to prepare for AI. According to Gartner, businesses rate AI as one of the hardest technologies to implement.
Business simply can’t afford to wait to figure it out – AI has gone mainstream.
When I look at the trajectory of AI adoption, I can’t help but remember the IT industry in the 90s. Back when the development scene was still relatively new in South Africa, I learnt to code using books in the library. Sure, coding existed, but it was a niche discipline that most people couldn’t even begin to understand. As more intuitive tools, learning resources, communities, and hardware became available, it became much more accessible and attractive to the general public.
We’re seeing that same transition take place with AI – the perfect storm of familiarity, data skills, infrastructure and software all advancing to a point where you no longer need a high level of specialisation to take advantage.
Platforms like Hadoop have been around long enough for an active ecosystem to spring up around them. The open source AI community is thriving, uploading new algorithms to libraries like Amazon Sagemaker, Azure Machine Learning Studio, and Google Cloud AutoML. Competitions such as Kaggle are incentivising the creation AI-powered solutions and rewarding creative applications. The days of building algorithms from scratch are behind us.
There are now a wide variety of APIs available that sit on top of the more baseline tools, allowing for greater complexity while lowering barriers to entry. Anyone with basic programming skills can take advantage of more intuitive deep learning platforms like MXNet and TensorFlow and deploy new models with a single click.
We’re also seeing the rise of Machine-Learning-as-a-Service models, where you no longer need to invest in a dedicated data science and AI team to access high-level machine learning. IBM Watson has its own as-a-service solution, which includes intuitive drag and drop interfaces that allow you to do things like model neural networks. Dell meanwhile has created ready bundles, which pre-integrate the hardware and software necessary for deep learning.
There’s no downplaying how much of a gamechanger all of this is. All of the new and more efficient operating models that have previous only been available to large global enterprises are now available to everyone. The democratisation of AI means that any organisation can potentially use data, deliver value, and restructure the workplace the way that Amazon, Tesla, Google and other giants are doing.
As with any major new technology, there’s a learning curve to overcome in welcoming AI. We may be getting there with the tools and platforms, but AI still represents a monumental shift in the way we live and work. Getting to grips with that world is about more than just learning to build an algorithm. The proliferation of resources and use cases, combined with greater adoption in both consumer-facing and enterprise environments, means we are effectively at Point Zero for learning to work with AI.
The first step is to understand exactly how AI can be used to creatively solve long-standing problems or augment human performance. Examples abound in every industry, across every continent of AI being used in practical but meaningful ways that don’t need Watson to execute. CLEVVA, for example, helps improve customer service staff’s decision-making abilities rather than replacing them, while Xineoh uses machine learning to predict customer behaviour.
There is a lot of sensationalism and misinformation around AI – thankfully there are a plethora of educational resources out there that can help demystify it, whatever level or industry one is in. Coursera, EdX, Udemy and Fast.ai are just a few of the platforms that offer foundational courses on AI and machine learning.
Now is also the time for organisations to start testing out new AI applications in pilot environments – and I’m not just talking about rolling out a chatbot and calling it a day. Think about existing challenges in your business and approach each on its own merits. Joining a platform like Zindi and Machine Learning Intelligence of Africa is a great way of keeping up with local and regional use cases of interest.
Preparing for the Fourth Industrial Revolution doesn’t mean replacing your entire workforce with robots tomorrow – it does mean thinking about working in new ways and embracing some of the advances of technologies as they become mainstream.