At the recent New York Infrastructure Masons event I attended, the application of Amelia Orchestra and similar technologies to support the Data Centre (DC) sector was a key topic of discussion, and great opportunity for industry leaders to openly engage around future adoption.
As economies become more reliant upon automated and machine-driven technologies to maintain and increase productivity, their general ability to create new jobs will likely become more limited1. The current prognosis is wages and the numbers of jobs available in classic industrial sectors will continue to decline, as automation and machines progressively find their place. For some time, digital infrastructure advocacy groups have been pondering over both the impact and application of increased AI and machines; this note reflects my personal take on the current state and platform for innovation.
When compared with the DC industry financial and health services are ahead of the curve when it comes to working AI models, and we can learn by engaging with early adopters in these sectors, around lessons learned and opportunities. Scaling AI solutions to support Edge distributed business models, and distributed large data volumes, is one of the current exciting initiatives, to allow a full customer experience at local retail level e.g. optician store trials in the US which are well underway.
Looking to the Benefits
Certainly, conversational AI, with its ability to talk to customers in unstructured language could be considered helpful to schedule and maintain change orders, to liaise with vendors and eliminate low level tasks for people. Database management across a single site or portfolios of sites through embedded use of conversational AI, provides space for redeploying and up-skilling the current workforce, and may help to alleviate critical resource shortages in other areas of DC operations.
So, what roles may be best candidates for augmentation or could be replaced by digital labour and what challenges might be faced through adoption of these technologies?
From an infrastructure investors’ viewpoint – general scheduling of tasks, billing, supply chain management, background security clearance, financial reconciliation, and general quality control are the obvious quick wins.
The ability to develop digital maintenance ‘whisper agents’ to shadow humans and develop a suite of learning tools to respond as site back up, is a wider resilience opportunity, as we look to harmonise with digital platforms and work collaboratively in tandem.
From an operator’s perspective – network load management, proactively readying and communicating at an enterprise level the need for expansion, typically a manual decision now, to enable better line of sight around lead times with vendors, will bring a positive level of efficiency, i.e. increasing the capacity of the network pipe, rack design and deployment, commissioning support and plant program management. SLA management from a reporting perspective, timely monitoring, generation of current state and incident reporting for customers are top of mind.
Adding to these benefits are supporting HR processes, addressing faults, the replacement of internal audit functions by providing a platform for continuous process improvement, and automated marketing functions – all opportunities where digital labour will help accelerate customer service and experiences. Think about trading floor applications, for example Bloomberg with circa 10,000 desk moves per year, and being able to free up the laborious human planning element and better focus this time on positive people engagement experiences –using AI to proactively plan, communicate changes with staff and vendors, swiftly and with apparent seamlessness.
Expect There to be Challenges
Widespread AI adoption may include regulatory / union considerations as personnel are potentially displaced and / or redeployed to new roles, and from a sensibility standpoint – conversational AI will only be as resourceful as its learning model. Developing the appropriate machine learning ‘smarts’, the learning algorithms to provide breadth and depth of knowledge to ensure customer experiences are memorable for the right reasons, and AI tasks optimised to at least the same standard as that of a human, will take ongoing development, testing and evolution.
Trust in the product is likely the biggest aspect to tackle – when many are wary of who they should trust, and have an inherent distrust of machines, this is one massive hurdle to overcome if unilateral adoption of AI and its benefits are to be tangible.
Machine Compliance & Regulation
But what if our unstructured AI doesn’t follow the so-called rules prescribed by us as humans, what if it becomes smarter and takes on an attitude and mind of its own?
Human error and deliberate sabotage accounts for a number of DC uptime failures and security incidents globally, on an annual basis. This is an inevitable outcome of human nature, but do we believe we can program maliciousness or rogue performance out of digital labour sources?
One question I ponder is the social responsibility of who designs and trains the AI – how exactly will they be trained? Is the end game about achieving more, faster and with less people to drive profits? For me, I think it’s about creating opportunities to balance innovation to advance and generate resilience in the sector, which we all now rely on so heavily to be 24/7 connected, while generating value for many through retraining or complementary roles which broaden the attractiveness of the industry.
Regulatory compliance will play a significant part in the appetite for adoption, with existing laws struggling to work through a mire of issues: ‘governments around the world are moving quickly to ensure existing laws and regulations remain relevant in the face of technology change and can deal with new, emerging challenges posed by AI.’2 In the 2021 [Cognilytica]3 report which explores the latest regulatory actions taken by countries around the world, its research around AI-relevant data privacy, conversational systems and chatbots, AI ethics and bias, AI-supported decision making and malicious use of AI are all highly relevant to the data centre sector. Findings highlight most countries are adopting a ‘wait and see’ approach to regulation and many countries still have very little or no laws and regulations regarding AI.
The consistency with which regulation is determined across geographic regions will be something to watch. As AI is more broadly adopted, operational risk, insurance considerations, cyber security and third-party reliance will all be considerations for board risk matrices and governance frameworks.
Speed of Adoption
With sector engagement wide amongst industry players, this suggests a good level of positivity around exploring the potential. Rapid digital adoption during Covid-19 has resulted in many organisations being more future focused and less resistant to change, setting a ripe platform for innovations in AI.
To me this is a possible scenario.
The opportunities for digital labour to support and enhance the DC sector are numerous and wide ranging, whilst in several respects are controversial.
Education, in conjunction with building trust, will be an enabler and catalyst to adoption. Communicating the benefits to reduce any threatening element to livelihoods through role displacement, investing in training to reskill and redeploy and demonstrating this happens, will be essential as the sector strives to build a new pipeline of DC talent and capability, and maintain trust by those affected.
The industry is about challenging design norms and striving for innovations which complement or add value to our rapidly changing digital landscape – we should be optimistic around how this can accelerate digital equality and provide opportunities.
Let’s celebrate the work underway, the human smarts behind development and the entrepreneurial investment in this space, but naturally, make up our own minds around how and whether we wish to leverage this innovation across our businesses, client, and customer ecosystems. I’m looking forward to this time next year and a new take on how much the sector has positively embraced AI throughout 2022.
World Economic Forum
Report: ‘Worldwide AI Laws and Regulations 2021’