The Environmental Cost of NOT Using AI

While activists debate AI's carbon footprint amongst themselves, the industries killing the planet are using it to stay ten steps ahead of us.

Artificial intelligence has the potential to make animal rights and environmental advocates orders of magnitude more effective than we have ever been. Yet, many advocates remain hesitant to embrace these technologies, often citing concerns about AI's environmental footprint. This reluctance isn't just based on misinformation, it's a gift we’re giving on a silver platter to the very industries we're trying to dismantle.

If we're serious about animal and earth liberation, we need to be using every tool available to win, including AI. Because while we're debating the ethics of using ChatGPT, factory farms are using the same technology to figure out how to make their operations even more efficient at exploiting animals and the planet we share.

The Real Environmental Impact of AI (Spoiler: It's Not What You Think)

Training and using AI does have an environmental impact, but so does literally everything else we do. The question is whether that impact is justified by the outcomes. Let's cut through the misinformation with some actual numbers:

• A single serving of beef emits 15.5 kilograms of CO2e.

• The ENTIRE training run for our Open Paws LLM emitted only an estimated 4.52 kg CO2e.

• This calculation is based on our 3.53-hour training time using 4x L40S GPUs consuming ~320W each (total 1.28 kW), resulting in 4.52 kWh total energy use, multiplied by ~1.0 kg CO2e/kWh for average data centre emissions.

That's less than 1/3rd the emissions of a single serving of beef (15.5kg CO2e).

Let that sink in. If this LLM convinces just one person to skip one burger, it's already carbon-negative. And that's before it helps us organise and automate campaigns, expose industry corruption, deliver personalised outreach to millions of people, or any of the thousand other ways it can advance our cause.

We kept our environmental costs low by using energy-efficient training methods. But even if we hadn't, the impact would still be a rounding error compared to what we're fighting against.

Putting AI's Impact in Perspective

While some cherry-picked studies highlight substantial carbon footprints from training the largest AI models, the reality for most advocacy organisations using existing AI tools is radically different:

• One ChatGPT query uses about 0.3 watt-hours – that's less electricity than your laptop uses while you're thinking about what to type in your next email to a potential donor

• Most studies show around a 60-150% productivity boost from using AI. Imagine doubling your organisation's output without hiring a single new person!

• Even if each query only saves you 5 minutes of time on your laptop, that’s actually enough to REDUCE your carbon footprint, rather than increase it

The math isn't complicated: when you factor in the net environmental effect, AI tools lead to an overall reduction in emissions by increasing efficiency. This is especially crucial for advocacy organisations that are already stretched thin trying to fight billion-dollar industries with the budget of a punk band's merch table.

Who Benefits When Advocates Don't Use AI? (Follow the Money)

It's worth asking: who stands to gain when animal and environmental advocates avoid using AI tools? Industries that rely on high-carbon activities certainly benefit when advocates ignore these powerful technologies, as it maintains their status quo, while they themselves leverage AI to maximise their efficiency and profits.

You'd have to be fooling yourself not to think the animal agriculture industry, fossil fuel companies, and other environmentally harmful sectors are already investing heavily in AI to:

• Optimise their productivity and profitability

• Predict and counter advocacy campaigns before they gain momentum

• Manipulate public opinion through targeted messaging

• Identify and exploit regulatory loopholes

When advocates choose not to use the same tools, they're essentially showing up to a gunfight with a strongly worded letter. These industries WANT you to fight inefficiently.

Every day we cling to outdated methods while our opposition embraces technological advancement is another day we've voluntarily handicapped our movement's potential impact. The status quo is killing trillions of animals and destroying our planet, and the status quo loves nothing more than when its opponents refuse to use the tools that could actually help them win.

Environmentally Responsible AI Use

Rather than avoiding AI altogether, advocates should focus on using these tools responsibly:

1. Choose Efficient Models: Not all AI tools have the same environmental footprint. Smaller, more efficient models can often perform advocacy-related tasks just as effectively as larger ones. You don't need a nuclear reactor to power a light bulb. More efficient models don’t just have less environmental impact, they also cost your organisation less money. And the smallest models can even be run locally on your laptop, keeping your sensitive data safe.

2. Consider Cloud Providers: Select cloud computing services that use renewable energy for their data centres. Vote with your dollars for a greener cloud infrastructure and run your AI in regions with lower carbon emissions.

Conclusion

The narrative that AI is environmentally harmful is not just wrong – it's actively harmful to our cause. It keeps us fighting 21st-century problems with 20th-century tools, while industries that harm animals use every technological advantage available.

The question isn't whether we can afford the environmental impact of using AI, it's whether we can afford NOT to use these tools when the stakes are so high. Every day we spend being less effective than we could be means more animals suffering, more habitat destroyed, and more entrenchment of systems we're fighting to change.

The animals and ecosystems we're fighting for deserve our willingness to use every tool at our disposal. They deserve advocates who fight smarter, not just harder. And they deserve a movement that's willing to embrace the tools that can actually help us win.