Recently, several of Cisco’s EMEAR innovation centers – namely those in London, Paris and Berlin – played host to a series of ‘AI days’. These events gathered subject matter experts with leading technologists from across the world, to discuss, display and showcase the power of artificial intelligence (AI) in transforming the enterprise and revolutionizing the way we do business today.
What emerged was a widespread recognition – by our EMEAR partners, customers and innovation centers – of the enormous opportunities inherent in AI, but also the challenges involved in adopting the technology on a wider scale.
AI today – the boom years
Over the last few years, interest in AI in EMEAR and elsewhere has exploded. In 2018 alone, UK VC investment in AI startups reached $1.3bn. The German Government has set aside €3bn to 2025 to fund its national AI strategy, and France has promised €1.5bn by 2022 for AI research and development. By 2022 worldwide spending on cognitive and AI systems is expected to reach a whopping $78 billion.
There’s a strong drive towards enabling our ‘AI future’ through education. Recently the UK government has pledged £110m to fund 1,000 places for students to complete PhDs at 16 Research and Innovation AI Centers across the country. And there are similar initiatives being repeated around the world. Cisco itself has joined a cohort of businesses – including Google DeepMind and BAE Systems — in funding 200 new AI masters courses at UK universities.
It’s clear why companies, Cisco included, support educational and knowledge-transfer initiatives such as these. If we can engage younger generations in the development of the technology, we’ll ensure its place in our future.
Europe is home to as many as 2,500 AI start-ups – many of which will spend vital time developing their solutions in one or more of Cisco’s innovation centers. Our ‘AI days’ are intended to inspire our customers and partners across the globe, and serve as an opportunity for innovation and knowledge transfer. Using our innovation centres as a platform, we aimed to share our unique approach to innovation and highlight the active role we’re playing in AI technology.
In these AI ‘boom years’, lots of people want to talk about the potential of AI. But for us, it’s important to show that we’re really adding meaningful value to the conversation – and that we actually have the capability to use AI to meet business outcomes.
Challenges to overcome
Over the course of Cisco’s ‘AI days’ it became apparent that there are some pain points for start-ups looking to adapt this technology for the first time.
Firstly, understanding what AI can and can’t do is key to success. And it isn’t always clear.
There’s a substantial gap between the capabilities claimed by AI companies, or discussed in the press, and what can be actually deployed at scale.
Anomaly detection is also a problem.
At first glance, it looks fairly straightforward: the AI processes the information it is given, and identifies a baseline. Then when it finds data points that don’t fit in with this baseline, it classifies them as outliers – far from normal.
But who decides what is normal and abnormal? There is no clear definition, because it so often depends on context. And AI isn’t so good at taking context into consideration.
Another challenge that businesses are facing involves the gap that exists between key stakeholders when it comes to realising the impact of machine learning (ML) and AI. CIOs see the competitive advantage that AI can offer, and place it top of their agenda. However, data scientists and data engineers face rapidly evolving open source ML frameworks and sometimes have a hard time scaling for production. CIOs look for use cases, resources and impact – sometimes more quickly than they can be achieved.
Also, enterprise IT teams today typically lack sufficient AI expertise. So, to face the AI implementation challenge, they will need new infrastructure architectures to solve silos and manageability issues.
So it’s clear that while AI presents a potentially powerful solution to many business issues, it also raises entirely new challenges.
An AI-ready approach
But while these challenges may seem daunting, they are far from insurmountable – as many of our partners and guests at the EMEAR ‘AI days’ demonstrated.
By bringing together the event attendees and showcasing all their different stories of success, it became possible to draw connections.
A pattern emerged that, whilst not exactly a formula for AI success, offered a rough philosophy to adopt. It started with a suggestion made by a customer from Berlin: when focusing on your AI strategy, you should “automate the easy – and augment the hard.”
This simple aphorism hints at another important aspect of AI strategy: to ensure that your use of AI is responsible.
Businesses must make their algorithms available for regular auditing to ensure that they comply with not only the law, but all of the other ethical considerations you need to bear in mind too.
Partnership is key to AI success
Our ‘AI days’ highlighted how far we’ve come with this revolutionary technology. But it also showed what we still have left to do.
As always, partnership will be essential. Businesses need to engage with experts to find the solution that’s right for them and abides to legal and ethical mandates.
The growth in AI technology presents a huge opportunity right now. But there are growing pains to overcome; and finding an appropriate partner, with the right technology solutions, will enable enterprises to seize the advantage that lies beyond the challenge.
As the AI days themselves showed: when everyone gets together and shares their understanding, we can reach better outcomes at a faster speed.