Artificial intelligence and machine learning have the potential profoundly to change lighting and lighting design. It is therefore imperative the industry and lighting professionals alike engage with it – and quickly – to understand fully its opportunities, limitations and, indeed, possible threats, argues Guus Ketelings.
Open any newspaper, listen to any podcast, turn the TV on, or even speak to people on the street and you’re very likely to hear the words ‘AI’ (artificial intelligence) or ‘ML’ (machine learning) at least once.
It’s certainly been the buzzword of the past couple years, particularly with the public release of OpenAI’s ChatGPT, Google’s Gemini and Microsoft/GitHub’s Copilot.
In the last decade, advancements in this field have started to permeate society rather than being isolated to research papers and academic trials.
Colloquially, when people refer to AI (and less so ML), they usually mean ‘Large Language Models’ (LLMs), of which a good example is ChatGPT. In the absolute simplest of terms, these models are capable of generating text based on prompts by predicting the statistically most-likely next word.
This can be very useful for quickly generating text concepts, reports, policies and other large swathes of text.
Some of you may also be using generative code, image, video and/or music AI – all of which functions in a more or less similar way to LLMs.
AI and ML are so much more than this though. Essentially, together, they encompass any processes that:
- Learn. By acquiring information and the rules that govern such information,
- Reason. By enforcing these rules and reaching approximate or definite conclusions,
- Self-correct. By checking the conclusions against a set of desired outputs (in other words the prompt, data, and so on).
ML is a subset of AI. It focuses primarily on developing algorithms and models that enable computers to perform tasks without instructions.
I have spoken to many people around the industry to probe them on their thoughts on AI and ML.
The typical response is that they have either dabbled with it or have heard of it but not yet experimented with it. A handful of colleagues I work with currently using it in limited capacity.
When asking about future potential, the responses boil down to the following:
- Using LLMs like ChatGPT and Gemini to automate report writing for lighting designers, lighting impact assessors, and engineers.
- Using LLMs to generate policies and contract documents for HR teams within the lighting industry.
- Using LLMs to support sales teams with bid preparation, engagement reminders and likelihood-to-buy analysis.
Of these three, it is the first point that is most often mentioned, in other words people are using AI as, essentially, a writing assistant. To me, this is very disappointing, as we are failing to truly unleash the power of such tools.
AI is already disrupting many industries. We in lighting should therefore ensure we are prepared to adopt this change – as an industry, organisations and individuals.
Preparation will likely require broad involvement from many teams. Organisations should be asking themselves – right now – how they are predicting to roll out new technology, map the inputs and outputs, train and upskill their staff and ensure other facets of the organisation is ready for such change (financial, HR and so on).
As an industry, I think it’s imperative that we make a concerted effort to bring software developers and algorithm experts on board – not in an advisory capacity but actually hire them into permanent roles.
This will give us the expertise we currently lack and desperately need to support us in this new digital transformation.
Finally, I’d argue the lighting industry would perhaps also benefit from a working group to scope out how we can support both employers and employees with this roll out. It could also offer guidance on best practice, recommendations for how to integrate AI into engineering and design processes and offer a platform for discussion via organising events and conferences. Maybe something for the ILP perhaps to consider taking a lead on?
Personally, I think that, on an individual level, keeping up to date, self-educating in low-code environments and monitoring other industries will give all of us a good foundation to adapt to this change.
Let’s embrace this technology fully so that we remain relevant and attractive to the next generation of lighting professionals.
Guus Ketelings AMILP EngTech is director at Let’s Light
*For those unaware of the term, ‘deus ex machina’ is Latin for ‘God from the machine’. It refers to a person or thing, or in this case potentially a technology, that is introduced into a situation suddenly or unexpectedly, and the consequences this can have.
This is an abridged version of the full article that appears in this month’s print edition of Lighting Journal. To access the full article, please click on the link to the edition.
