What's the Environmental & Ethical Impact of AI?
Artificial intelligence has been with us for several decades, but it’s really starting to make its presence felt of late. You can hardly turn anywhere these days without encountering it.
AI is used in many industries, including healthcare, transportation, agriculture, conservation, education, and finance.
It has led to many exciting developments, but as with most things the pros come with cons.
How AI Can Be Used for Good
One of the great pros is that during disaster relief, AI models can assess damaged infrastructure. This saves time for first responders to focus on rescue efforts.
The use of AI models in the medical field can lead to the faster and more accurate detection of cancer. The development of autonomous vehicles may result in fewer injuries, traffic jams, and carbon emissions.
However, the latter can be countered by the fact that these vehicles require vast amounts of resources to manufacture.
AI often presents such conundrums, as its benefits are often countered by drawbacks. These drawbacks can often be overlooked, at times even hidden.
Negative Impacts of AI
It was recently been reported that getting a ChatGPT search answer requires nearly 10 times as much electricity as a typical Google search query. Let’s explore some of the negative environmental impacts of AI.
AI Training Resources
The training of AI models involves the processing of countless data points, images, and algorithms, consuming hundreds and even thousands of megawatt hours of electricity, emitting hundreds of tons of carbon as a result.
AI model training requires a lot of water, due to onsite server cooling and offsite electricity generation. This places added stress to freshwater resources, especially in areas where this is already limited.
The consumption of resources may not be as intensive during the running and maintenance of artificial intelligence, but it does not stop either.
AI and the Fossil Fuel Industry
Electricity used to power AI is often derived from burning non-renewable resources like coal, oil, and gas.
These are fossil fuels that bring with them a range of detrimental environmental impacts like resource depletion and greenhouse gas emissions when fossil fuels are extracted from the Earth.
Emissions continue pouring into the atmosphere when fossil fuels are burned to generate electricity.
The manufacture of hardware must be considered as well. Not only does their manufacture require energy consumption, but it is also tied to mining demands.
Elements that are needed to build AI infrastructure – such as copper and lithium – need to be mined, and the mining industry comes with its own plethora of environmental, as well as humanitarian, issues.
While there is a push for the AI sector to seek more sustainable means of running, and especially training, models by relying on renewable energy sources, it feels like these goals are rendered moot by coal, oil, and gas companies investing in AI to seek and exploit more reserves.
And the demand for AI is only expected to grow, increasing the demand for electricity, expected to double by 2026. This growth has already resulted in an increase in the number and size of data centers.
Data Centers
Data centers – where you will find servers, networking equipment, and storage systems – house and run AI. These consume a lot of electricity, usually running around the clock, supplying an uninterrupted flow of power for backup purposes.
Furthermore, space is needed to build data centers. To facilitate this, land is cleared, leading to habitat destruction.
The presence of data centers in urban areas is also a concern, as it can impact the infrastructure of surrounding communities. The high demand for electricity can increase grid congestion, raising the price of electricity for those in the area.
unethical AI Biases
When AI models prioritize economic growth inputs over environmental protection or social responsibility, it can lead to less sustainable and ethical outputs.
Unfortunately this is a very human problem. The information that is inputted into AI models comes from humans, which means there is often an inherent bias built into the system, reflecting historical and existing biases not only of putting profit over people and the environment, but also biases of race, class, and gender, further disempowering certain groups.
Furthermore, the proliferation of AI in society can encourage overconsumption, further impacting the environment.
The use of AI technologies increases the efficiency of supply: products become cheaper (but not necessarily of better or even the same quality) and more accessible (think of today’s fast and frequent deliveries).
E-Waste
End of life issues are also a problem for AI. When devices become obsolete, they are disposed of. If this final step is not managed sustainably and responsibly, the negative environmental impact is exacerbated.
Since hardware is constantly being advanced and upgraded, old components are quickly rendered obsolete. This results in electronic waste, which contains hazardous materials like lead, mercury, and flame retardants.
Pin this post for later!
When these enter the environment, they can contaminate soil, air, and water. AI’s e-waste needs to be responsibly managed, either through proper disposal or by recycling parts or extracting those that can be reused.
Quality must be prioritized. It is one thing to improve a device, quite another to plan for a device to become obsolete only to generate further sales. This is known as planned obsolescence.
In an effort to curb this strategy, there is an effort to hold manufacturers accountable through policies like Extended Producer Responsibility. This policy holds manufacturers accountable for the whole lifecycle of their products, including afterlife management.
This can take place in the form of proper disposal, recycling, or take-back programs that allow them to reuse parts.
Can we find a balance between good and bad AI for the sake of sustainability?
AI models can have positive outcomes for the environment too. Food loss and waste are a large issue for the environment.
In the agricultural sector, AI has been used to curb food loss and waste by predicting consumer demand and optimizing the use of resources. Similarly, AI is being used to monitor conservation efforts.
However, this can have negative impacts too, like when AI systems such as drones disturb wildlife habitats and migration patterns.
In the mining sector, AI models could help find alternative sources of lithium: an element that is crucial for the development of green energy.
In an ironic twist, lithium mining comes with a heavy social and environmental footprint. By implementing new AI technology, extraction can hopefully become more sustainable.
There is a great but challenging need to strike a balance between reaping the benefits and reducing the risks, especially since AI’s negative environmental impacts cannot be ignored in the age of the Anthropocene.
The AI sector needs to make more concerted efforts to be sustainable. This is particularly important in the resource-intensive training phase, as well as in constructing data centers in a way that is less disruptive both to people and the environment.
Solutions have been put forward such as using renewable energy, developing more efficient hardware and algorithms, building devices to last, and shaping AI policy in ways that take the environment and social responsibility into account.
After all, the world’s best computing device remains the human brain, so it is up to us to make anything that resembles it equally efficient.
MEET THE AUTHOR
Claudia Hauter is a South African writer, copy editor, and content creator with degrees in Drama and Anthropology. She works in television managing web content. When she isn’t reading or writing, she’s walking her dog, finding small businesses and markets to support, or attracting butterflies and bees with her vegetable garden. Learn from Claudia on Twitter or Instagram.
More Blog Posts on Conscious Living:
For more ethical & sustainable tips be sure to explore more conscious lifestyle blog posts, follow The Honest Consumer on social media, subscribe to our newsletter, & check out the Ethical & Sustainable Brand Directory.