Tim Fleet, Vice-President of Business Development at Idox
Generative Artificial Intelligence (AI)s such as ChatGPT are well known for their ability to enhance human creativity by making realistic variations of existing content from auto-generated emails to whole books co-authored by machine. However, few realise that a recent paradigm shift in AI could also help overcome a growing energy skills gap by democratising decades of knowledge on everything from energy design to maintenance.
Today, vital energy industry knowledge is often fragmented and accessible only to specialists. This hinders learning and development and prevents energy workers from accessing important information that could hold the key to improving infrastructure development or transforming operational safety. When data is sequestered behind departmental boundaries, this also hinders multidisciplinary collaboration and creates one-dimensional teams. Crucially, vital skills held by a handful of workers can be lost forever to retirements or layoffs, exacerbating skills shortages.
Just as programming languages once helped translate human instructions for machines, generative AI now holds the promise of translating specialist information for general human consumption, democratising skills across the energy industry. By rapidly synthesising, summarising and translating millions of records for human consumption and unlocking new insights from them, AI could also accelerate innovation and supercharge the green energy transition.
The skills bottleneck in the energy transition
We are witnessing a major surge in demand for energy infrastructure with more new renewable energy capacity to be added in the next five years than in the previous 100. This is taking place alongside soaring demand for low-carbon fuels with a projected surge in Liquified Natural Gas (LNG) projects and nuclear power construction.
Yet the boom in new energy projects is being hindered by an unprecedented skills shortage. This is driven by a perfect storm of an ageing workforce, demand for new skillsets to support digitalisation and decarbonisation, and the misalignment between traditional energy education and the skills needed for new energy.
There is an urgent imperative for energy skills and knowledge to be preserved and passed onto a new generation and for workers to be rapidly upskilled in the energy systems of the future. New advances in Large Language Models now have the potential to help overcome these skills bottlenecks and democratise energy knowledge and skills.
The emerging paradigm in human-computer relations
Human-machine interactions have undergone many paradigm shifts over time, from the days when human instructions were translated for machines through paper punch cards to the rise of voice user interfaces. LLMs represent a similarly game-changing evolution that could enable machines to return the favour by translating complex data for humans. These innovations enable machines to turn huge amounts of complex data into human-like text in any format or language, transforming human-machine interactions from one-way commands into two-way conversations. This has the potential to break down knowledge silos and turn data into a resource accessible to all workers.
Crucially, many of the risks associated with AI are now being overcome. While platforms like ChatGPT have fallen prey to errors and unreliable data due to the fact they were trained on data from the internet with no quality checks, industry-specific LLMs are changing the equation. LLMs can now be trained on reliable, industry-specific data such as verified energy safety records, operational best practices and regulations. They can then rapidly synthesise and summarise this data to auto-generate multilingual training materials drawing on the latest lessons learned, regulations and best practices.
LLMs could even scour records to identify the biggest skills deficits and deficiencies across an organisation and amass the leading expertise in those areas, thus helping both find and fill skills gaps. In future, AIs could even suggest novel new forms of training or improvements to everything from design to operations.
Their capacity to rapidly amalgamate and translate millions of records for human consumption means that LLMs can also fuel cross-departmental knowledge sharing and help create a more multidisciplinary workforce. This would put enormous cognitive resources at the fingertips of workers and could unearth vital new insights into everything from energy operations to design. For example, workers could ask an LLM to identify and explain the three most common design faults across wind farms since they were first built, helping develop smarter future wind designs. Likewise, workers could ask an LLM to identify common faults affecting each battery type to help optimise battery usage.
As well as making information more widely accessible, AI could also democratise innovation itself. By deriving new innovations from patterns in existing data, LLMs could, in the future, auto-generate project templates or even infrastructure designs.
Data is the key
The success of AI technologies, though, depends on a number of factors. Key to the endeavour is ensuring the quality and availability of industry data. Engineering information management systems already widely used in the energy industry can now automate processes like version control and enforce rigorous document management standards such as sending reminders of overdue document deliverables.
For example, engineering information management systems now enable all data to be stored in a central digital environment with a complete audit trail of all changes and easily accessed through tag-centric search functions. Automatic revision control and auto-generated version histories of all changes further ensure the data integrity required for LLMs. Together, these systems are helping prune and preserve energy industry data in pristine condition for future AI applications.
This could transform industry knowledge and skills into a globally accessible resource that could be unlocked for new generations of workers. The resulting democratisation of information could transform workforce training, ‘level up’ knowledge across organisations, and unlock some of the biggest skills bottlenecks in the energy transition. It also holds the key to accelerating the energy transition itself by helping the industry realise the full potential of its immense data resources.