Opening up new AI applications

Opening up and demystifying artificial intelligence and machine learning is the best way to overcome perceptions that AI is a threat to people’s jobs

There seems no doubt that the open-systems, open-platform approach through the latter part of the last century liberated IT from its once proprietary and lofty perch, ushered in the universal connectivity of the internet, and put technology into the hands and pockets of everybody.

This is a good thing, not because of altruism, but fundamentally it stimulates and encourages innovation. Technology is one place where “levelling up” can be seen in action producing good results. Programming has not for a long time been the preserve of a select few assembly language programmers crafting the control systems of Apollo moon shots, but anyone, anywhere can visually or naturally create and combine logic to generate new outcomes.

The compute, connectivity and capacity power of technology enables it to do more of the digital mundane, presenting people with an ever-higher level of abstraction and empowering individuals to be more creative and productive. While it sometimes seems that new technology will simply “replace” people, employed properly, it can augment and enhance human endeavours, as any effective tool should.

Why, then, does artificial intelligence (AI) so often get portrayed as something far more insidious and threatening?

Part of the problem is the mystique. The IT industry, like many others, is prone to build up hype, which can portray some technology as so special that it needs to be restricted to a certain group of users, imbued with almost magical powers. Although specialist skills and training are vital, especially in the early days, most innovation only really blooms when new technology moves into mass adoption and use, often with surprising applications and unexpected consequences.

AI has been through a long gestation phase, but with machine learning applications in particular now being much better understood, it seems ripe for wider use. Supporting this requires the increased accessibility and diversity that comes from ease of use, but also from novel examples and use cases, and reduced financial barriers to entry. Increased usage breeds familiarity, then acceptance and understanding of what it can (and can’t) do and where best to apply.

Driving this process of “democratising” AI involves creating communities of trusted partners to make a positive impact. Flexible technology models, such as the use of open source software and the delivery of capabilities dynamically – such as AI as a service – form part of the process, but alone are insufficient. It also requires a fundamental attitude shift towards ease and speed of adoption and the willingness to build on, and combine, the expertise of others, to deliver easily interpreted results that add clear value.

This democratisation process is not easy for some companies because it involves significant give, as well as take. One company on a mission to make AI accessible to everyone is the open source software company, headquartered in Mountain View, California. While its workforce is heavily oriented around highly experienced data scientists, it understands the mantra that “data is a team sport” and increasingly relies on a broader cohort of roles.

To be properly understood, AI needs to shift away from being technical discussions about data, machine learning and complex interpretation, to more about stories or business use-cases with immediate applicability.

Read more about opening up AI

Some verticals are frequently the earlier adopters, but as the technology opens up, patterns change.’s financial services customers have already applied AI to fraud detection, insurance underwriting and predicting customer actions, but right now, when business models are more uncertain and difficult to plan than ever, many processes in all types of enterprise would benefit from a little guidance.

Healthcare seems an obvious place to apply AI and get results quickly, for example to assist with staffing predictions, modelling virus spread and consequent demands on facilities. But recent step-changes in human behaviour, from shopping habits to working from home, make planning for logistics, supply and consumption more complex and introduce too many new variables.

It is here where a more open model of AI, able to encompass new models and multiple external data sets, and bring together diverse data science and domain expertise, can make a difference.

Along with many companies in the evolving AI sector, has built its success so far on targeting large organisations and data scientists, leading to a plethora of machine learning models and information dashboards. The next step in the democratisation process is the story of AI being told in the heart of businesses large and small, with intuitive and extensible AI-enabled services being applied to enhance business processes.

Guidance with some tasks coming from real-time analysis creating actionable advice, and automation of the most mundane activities, allowing more time for the application of the human mind, is where it can add most value.

AI is not about humans being replaced by super-smart machines, but being better supported by them. Ironically, the most important thing about AI is getting more humans using their own intelligence to apply it to an increasingly diverse range of applications. Open and business-oriented AI seems to be the way to go.

Read more on Artificial intelligence, automation and robotics

Data Center
Data Management