As issues such as worsening weather patterns threaten food supplies, biodiversity loss is increasing at a rapid rate, and food waste is a large environmental issue, it stands to reason that many solutions should be considered when deciding how to address these problems. AI, with its wide-ranging abilities to detect and analyze patterns and predict future outcomes, is a tool many believe will be significant in solving these food system challenges. However, there are still ethical questions to consider before deciding to employ it in personal and workplace settings.

Many food businesses nowadays have sustainability as an aspect of their mission: whether it’s developing systems to reduce food waste, ethical and local sourcing of ingredients, or contributing to biodiversity. AI is an amazing tool and is already being employed to help address a myriad of environmental and food system issues. It’s also being applied more and more in many food-related businesses. In one recent survey, it was found that 41% of food retailers and 69% of food suppliers are using AI for everything from “optimizing decision making around pricing and promotion” to “hiring purposes.” However, given the well-known negative environmental impact and ethical questions, its responsible use within the food system is imperative. Understanding these negative implications of the use of AI will aid in making critical decisions when it comes to deciding how or even whether to implement it.

The use of generative AI negatively affects the environment in different ways, including by creating electronic waste and the use of rare earth metals in microchips. However, two of the most notable harms are the high water and energy use required to power and cool down the servers housed in data centers. One recent study by researchers at the University of California, Riverside found that one 100-word email output by ChatGPT is equivalent to pouring more than a 16.9oz bottle of water down the drain, and can be almost three times that amount depending on where in the country (or world) that search is being executed. This is shocking, especially with the reality that most people who use ChatGPT often complete multiple searches resulting in many paragraphs of generated text or images. While there is still significant research to be done on the waste and environmental effects caused by different AIs, this study helps us understand that the more complex AI applications used by many food businesses are likely to require at least the same, if not significantly more energy, than a simple text output does. It’s clear that the environmental toll can become quite high quite quickly. According to one paper, the projected water use of AI by 2027 will be the same amount as the needs of half of the UK. This is at a time when many people in the world still regularly lack access to clean water. 

The UC Riverside study goes on to say that if 16 million people (about 1 in 10 working Americans) use ChatGPT for a 100-word email output one time each week, which is a modest estimate of the amount that people use ChatGPT to begin with, it would equate to the electricity demands of all the households in Washington D.C. for 20 days. Additionally, the International Energy Agency reported that one search with ChatGPT is equivalent to the electricity needs of 10 Google searches. It’s important for a food business potentially deciding whether to integrate AI applications if they are willing to take responsibility for this water and energy usage and if the harmful environmental impact is worth the outcome. The answer to the latter is likely to be different depending on the purpose of the AI being implemented. For example, using AI to significantly reduce food waste across a large company could have a net beneficial impact environmentally due to the prevention of water, energy, and pesticide output needed to grow that food, as opposed to, say, an AI application that is focused on hiring and promotional purposes within the company or applications that analyze user data to individually promote unhealthy and environmentally-unfriendly products.

While several companies who have developed these technologies have made statements committing to reducing energy and water demand, currently no concrete targets have been made. The reality is that even if things can improve in the future, we still aren’t there yet. As AI usage continues to grow, the harmful environmental effects will increase. While this post won’t dive into the abundance of other ethical questions that AI brings such as biased algorithms and potential copyright issues by using the work of authors and artists without their knowledge, these are just as important for digital businesses, including food companies, to be aware of and consider before using AI. 

The use of AI can be very beneficial in the workplace and within innovation to potentially solve current issues facing society related to the food industry, such as loss of biodiversity, food insecurity, and increasingly extreme weather events. It’s also ethically critical for businesses to consider whether the negative environmental and human impact of AI will be worth the benefits. As the Chief Digital Officer of the United Nations Environment Programme (UNEP), Golestan (Sally) Radwan, emphasizes, “We need to make sure the net effect of AI on the planet is positive before we deploy the technology at scale.” For businesses within the food industry who often make claims of sustainability, sometimes largely due to customer expectations, this notion will require them to think critically and to fully understand their goals and the impact they expect to see before integrating certain AIs. Businesses must also continue to learn and stay up-to-date with emerging environmental and ethics research surrounding AI.

The following are simple questions to ask before deciding to use AI, whether on a personal level or by those making decisions within food companies. It is not an exhaustive list and many more could be added or customized depending on the user. The point isn’t to completely prevent the use of AI within food businesses and the food system given its potential benefits to solve real issues, but perhaps the questions will help prevent the unnecessary uses of AI in favor of other more ethical practices.

1. Why do I want to use this application? Is it actually necessary? 
2. If AI didn’t exist, how would I go about finding this information/solving this problem? What other options are available to me/the company?
3. Is this use worth the X amount of water and energy it will consume? What are the harmful effects of this use of AI?
4. Is the impact outweighing the harm?
5. Who is benefitting from this use of AI? Who would benefit if I solved this problem a different way?
6. Do I trust this company and the software/algorithm they’ve created?

A few general ways to be responsible with AI

• Inform yourself about the environmental impacts and stay aware of changes. Make sure this information is shared and known across the company so others can make informed decisions.
• Reduce the unnecessary use of AI. For example,
  • Within the food industry – Are AI hiring and human resource tools actually necessary?
  • When needing images – Creating an image with AI consumes much more waste than a text output and it’s very likely that an already existing image by a photographer or artist fits your needs, which would support a real person. Consider using stock photo websites like Unsplash which has thousands of free and premium photos.
  • Turn off unwanted and unnecessary use of AI in your daily life – check the apps that you use daily, such as search engines, health trackers, word processors, etc., and decide if their built-in AI is necessary for your needs. 
• Take a few moments before rushing to solve a problem with AI to work through the questions above and potentially find another method.

Businesses working within the food industry often have a unique and privileged perspective of current environmental challenges due to the interactions between environmental health and food production. It’s no longer acceptable to turn away from the problems that the food industry contributes to and is affected by, such as topsoil loss, contaminated water sources, and species death, in the name of cheaper food, “improved customer experience”, and convenience. AI can be a beneficial tool, but only when applied purposefully with clear net positive impacts. Food companies must take responsibility for the impacts caused by their use of AI and make bold decisions to reject unnecessary uses if they want to maintain respect in a world facing devastating and worsening environmental issues.

Leah Michelsen

Image source: Photo by Luke Michael, Unsplash