Modern arboriculture in an AI era

Callum Throw, Arb Innovators Ltd

At the beginning of September, I was tasked with leading a workshop at the Arboricultural Association Conference.

The topic I chose was not specifically related to arboriculture; I wasn’t talking BS (though some who came to the sessions may disagree) or discussing the latest breakthroughs in tree biology or risk management. Instead, I decided to explore the area of large language models (LLMs), specifically ChatGPT. Yes, I’m aware there are other LLMs, but this one is widely known and frequently misunderstood.

Much to my surprise, around 80 participants attended the workshop over two days, and guess what? They were not all under 30. Rather pleasantly, it was a diverse group of the excited, fascinated, perplexed, bewildered and sceptical – ideal!

I started the workshop by asking a simple question: ‘How many of you have used a GenAI (Generative AI) tool like ChatGPT?’ One or two attendees shot up their hands, claiming to be ‘experts’ and ready to join me in leading the session. Another couple gave themselves a pat on the back for being pretty darn good, while the majority of attendees claimed they’d never used a GenAI tool like ChatGPT or had used it infrequently but didn’t fully understand what a GenAI tool was for.

This wasn’t all that surprising. Despite its popularity as the fastest-growing app in history with 100 million users in just two months, ChatGPT use is still far from widespread, with less than 10% of respondents in a recent YouGov survey using it daily.

Here’s where it gets interesting

I divided the room up, mixed the ‘experts’ with the newbies and tasked them with answering numerous basic questions about the fundamentals of ChatGPT. I included questions on the model and how it works, such as ‘What’s the knowledge cutoff for ChatGPT?’ – some informed yet wild guessing occurred. ‘What’s a context window?’ – blank stares. When it came to tokenisation it was as if I had asked the group to explain the theory of quantum mechanics. You see, this basic understanding of how LLMs like ChatGPT function is integral to getting the most out of them.

A further follow-up question revealed that the majority of those in attendance utilised ChatGPT sporadically and for the obvious tasks, such as social media copy, rudimentary website or blog copywriting and the odd question and answer here and there. A common trend surfaced across the group: many had developed an entrenched practice of asking for an answer and accepting whatever solution was offered – a ‘that will do’ mentality similar to our ‘Google it’ mindset. The overwhelming majority of attendees accepted that most of us are unaware that AI should be your all-in-one tool for research, diagnostics, data analysis and even strategic planning.

Now, I am aware of the sceptics. I can almost hear the quiet murmurs as you read this article. I’m fine with it. You see, I appreciate that you can’t go a day without hearing someone talk about AI, seeing an advertisement for the next AI-powered assistant, coming across a social media post with the words ‘game-changing’, American spelling or the notorious rocket emoji. I accept that it’s easy to see why people may misinterpret individuals shouting about GenAI as overzealous and eager nerds quick to jump on the latest fad.

Trust me, this isn’t a fad. Simply put, AI and GenAI isn’t the future of arboriculture – it’s the present.

And one thing I’ll die on my hill for is this: if you’re not currently harnessing its power and incorporating it into your everyday life or business activities, you’re not just playing catch-up; you’re not even playing the game. This incredible technology is perhaps the single biggest advancement since the dawn of the internet and it’s here now. It is altering industries across the board. Arboriculture isn’t the exception.

Mindset, mindset, mindset

Over the past 18–24 months, the rise of GenAI as an accessible technology has revolutionised business processes and is already profoundly reshaping day-to-day tasks. According to a recent poll by Forbes, 54% of company owners and executives say that AI solutions have increased productivity in their businesses, while some experts claim that GenAI has the potential to automate up to 70% of employees’ daily activities. Others predict this number to be even higher.

That’s why a change in mindset is critical. LLMs are not just another search engine like Google (as I’ve mentioned – information gathering, one input to one output) or a tool you use when you are feeling lazy and looking for shortcuts. Forget using ChatGPT for low-hanging fruit for a second. Sure, it can do it all day long, but that’s not what I’m talking about here.

You have to start thinking of LLMs and GenAI tools as the powerful, intelligent and extremely capable business operating systems (BOS) they are, and recognise that they are already transforming how we approach many aspects of arboriculture.

How AI tools can streamline processes

The greatest productivity increases are found in knowledge work. If you sit in front of a computer carrying out similar tasks from one week to the next, you need to start integrating LLMs and GenAI tools so you can begin automating processes such as manual data entry, administrative activities, data analysis, report preparation, marketing and sales. All these tasks can be completed competently by GenAI, and perhaps more importantly, by individuals who use these tools to aid their work.

Report generation: By analysing data from field surveys and other sources, GenAI can streamline the process of creating comprehensive, accurate arboricultural reports that highlight key findings, insights and even recommendations.

Data analysis: A lot of day-to-day arboricultural consultancy involves handling large amounts of data from tree surveys. LLMs can analyse this data quickly and accurately, identifying patterns and trends that might be missed. This can lead to more informed decision-making.

Problem solving: GenAI can help arboricultural consultants generate and research ideas, explore a range of scenarios, and critically evaluate any potential solutions.

Content creation: Whether it’s writing search-engine-optimised (SEO) blog posts, social media updates or detailed articles, LLMs can not only produce high-quality text that resonates with your audience, detailed infographics, charts and tables, but tools like Midjourney or Dall-E can generate realistic images. This is particularly useful for SMEs looking to enhance their online presence without dedicating their already limited resources to marketing efforts.

Quickfire dos and don’ts

Now, before you go from dipping your toes in the pool to diving off the top board, here are some quickfire dos and don’ts.

Above all else, just like I told those who attended the workshop, avoid anyone selling you copy-and-paste prompts. We’ve all seen them on LinkedIn or Facebook, ‘Follow me, like and share and I’ll send you 100 game-changing prompts’. I’ll say it loud: these completely miss the mark. Don’t get me wrong, we’ve all been tempted to use those generic copy-and-paste prompts floating around the internet. Yes, they’re better than nothing but here’s the thing: they simply don’t yield great results.

Why? Because GenAI is generative, a prompt which works for someone else is unlikely to yield the same results for you. GenAI and LLMs thrive on context and specificity. These one-off prompts often lack the nuance of your specific situation, making the output as generic as the input. Whilst Cs get degrees, we’re after As so it’s time to put in a little more effort.

Navigating AI integration

So, you’re ready to go, right? Whilst the enormous potential of GenAI is widely accepted, the shift to adopt it in your day-to-day workflow necessitates careful planning. A good place to start is with the following tasks:

• Identify your needs: Determine which areas of your business could benefit most from AI. Is it data analysis, technical report writing, sales or marketing?
• Choose the right tools: Research and select AI tools that best meet your needs and financial budgets. Tools like ChatGPT, Claude and Perplexity offer different strengths and capabilities.
• Train your staff: Ensure that your team is comfortable using these tools. This will almost certainly involve training sessions or workshops.
• Integrate with existing workflows: Start slow. Integrate AI into your existing processes gradually to allow for a smoother transition.

Also, a huge word of warning! As we embrace AI, it’s vital to stay on the right side of privacy laws. Handling personal data requires strict adherence to UK GDPR regulations. Remember, when using AI, avoid inputting confidential or identifiable information to safeguard privacy. UK law is also still catching up with AI advancements and it’s currently a little unclear as to who owns AI-generated work, so proceed with caution and don’t rely solely on GenAI for content creation. Instead, use it as a collaborative tool, enhancing your work rather than replacing your expertise.

Lastly, while AI can be a powerful ally, blind trust is a no-no. Always verify critical information and use your judgement. AI is here to assist, not to dictate.

Your people remain key

Once you’ve identified your needs and which workflows would benefit from GenAI tools, it’s critical to openly discuss your findings with your teams and develop a structured plan to integrate AI and train your staff.

One thing I’ll die on my hill for is this: if you’re not currently harnessing its power and incorporating it into your everyday life or business activities, you’re not just playing catch-up; you’re not even playing the game.

I urge you to ensure that your employees are on board with this technological evolution and don’t feel left behind.

Crucially, training should focus on maintaining skills like critical thinking, creativity and problem-solving, whilst also emphasising the need for ethical decision-making and caution with sensitive data. This is paramount and cannot be stressed enough.

You should look to obtain valuable guidance from reputable sources like the EU AI Act and the National Institute of Standards and Technology.

Once you’ve adopted AI, you must regularly audit your AI systems to check for biases and inaccuracies to help protect both your business and your clients from potential risks associated with AI misuse.

We value your input

As the role of AI and GenAI in arboriculture grows, so does the conversation around its impact. So now it’s your turn to weigh in. I’d love it if you’d spare a few minutes to help me and the Association gather some valuable information by answering the following questions:

• What risks or challenges do you see with increased use of AI in arboriculture? Are there potential pitfalls we should be aware of as we integrate this technology into our work?
• What benefits or opportunities could AI unlock for our industry?
• How do you think AI could enhance your day-to-day operations or long-term goals in arboriculture?
• How can the Arboricultural Association support with this? Would specific training programmes, workshops or resources help you gain confidence in using AI?
• Should the Association update its Code of Conduct and Ethics to address the use of GenAI/AI? How do you think ethical guidelines should evolve to ensure responsible and fair use?

Join the discussion by scanning the QR code. Your insights and experiences will be invaluable in moving this conversation forward and shaping some future technical updates from the Association.

Images generated using Dall-E

Callum Throw is the Managing Director and Principal Consultant of Arb Innovators Ltd. Callum describes himself as ‘not a typical arb consultant’ and is dedicated to integrating his love of arboriculture with his passion for accessible technology, such as drones, photogrammetry and generative AI.

This article was taken from Issue 207 Winter 2024 of the ARB Magazine, which is available to view free to members by simply logging in to the website and viewing your profile area.