Artificial Intelligence (AI) has developed so quickly over the last few years that it has now permeated almost every facet of life. Even if we as consumers do not want it, businesses are trying to push it under our noses everywhere - often without an opt-out choice.

The problem is that AI is extremely costly. Globally, it eats up more than double the electricity that our country produces. Each individual query might seem small, but multiplied by billions of users, when it is unregulated, we end up with an environmental cost that is simply staggering. Growing numbers of data centres needed to support AI, often created with little public input, are having adverse effects on local communities - all to feed a technology that many of us feel we don't want and have little say in. In the US, the explosion of data centres have contributed towards an almost $60 billion of rate increases. Each data centre can drink about as much water as a small town. Local residents in Memphis, Tennessee are intending to sue because of the projected worsening of air quality due to data centres. Immense vibrations and noise pollutions are also in danger of harming local wildlife.

Is this the future we want for NZ? We call ourselves a clean, green country. We fought hard to remove nuclear energy from our shores; we can apply the same strength to AI.

Harming our communities and hurting our environment isn't the only negative of AI. The list goes on:

• MRI evidence is coming out that frequently using AI can harm our thinking skills. Like any ability, when we use critical thinking less, the skill atrophies. People argue AI is the same as using a calculator...but does a calculator tell you what to think, what to like, what to want?
• Data privacy has reached a new level of danger. Horrific new AI glasses allow tech companies access into our private lives. Your face is no longer yours: anybody can recreate it in generated photos and videos any purpose they like. And if a totalitarian government comes into power, they will have immediate access to every piece of information about you.
• Sexual harassment has reached new levels. Now, explicit videos of anybody - you, your work colleagues, your children - can be created at the touch of a button. This content destroys lives and careers, and the tech bros don't particularly care about stopping it.
• Low-level positions are increasingly being replaced with AI. Soon, our youth will have no way to enter the workforce.
• Mass unemployment. Companies love AI because it gives them a slave that can do anything immediately without being paid, will never get tired, and will never report sexual harassment. Even the NZ government plans to replace 9000 public sector jobs with AI in the next four years. But if nobody has a job, who will have money to buy these companies’ products?
• Stealing artists' copyright works. AI companies claim that it ‘transforms’ these works, but it often memorises and produces recognisable existing works.
• Convincing people to kill themselves.
• The widespread use of AI in hiring has been shown to have strong racist​, sexist and ageist biases - and are much harder to challenge legally. Studies show that even having these decisions reviewed by a human will simply increase the bias of the human too.
• Workers in AI supply chains labelling data describe their conditions as modern slavery.
• AI can and does make mistakes. It's becoming more common for people with real technical know-how to be fired because 'you no longer need technical skills with AI'. But when actual technical skills are undervalued, who will be able to check its work? Do we want this amongst our doctors, our engineers?

Can AI also have wonderful applications? Of course. But each time it is used, we need to think if the purpose justifies the costs. For medical research? Absolutely. For giving you erroneous facts every time you do a Google search? No.

Many major AI company CEOs openly admit that AI has a decent chance of destroying humanity, on the same level as a pandemic or nuclear weapons.

This petition does not intend to ban AI. It is simply a call that we need our regulations to catch up with technological development - and fast.

We propose some or all of the following suggested regulations to be implemented:

• All AI features must be done on an opt-in model. By default, AI features should not be enabled. This is similar to what the EU has enforced.
• Copyright work should be excluded from AI training sets. Using copyright work in an AI model is a breach of copyright.
• AI-generated content, or content created with the help of AI, including text, photos and videos, must be clearly labelled as such. Companies must disclose when using AI as part of goods or services.
• Everybody should own the copyright of their own body. Creating an AI-generated photo or video of somebody without permission will be a breach punishable by law. Recording a person or the interior of a private dwelling for the explicit purpose of AI training without permission shall also be prohibited.
• All AI features should have a warning sign explaining the amount of clean water that each search uses, cautioning that it will contribute to energy prices rising, and asking the user to consider if they need AI, in the same vein as cigarette packets.
• New Zealand should work towards global measures to mitigate the extreme frequency of AI use, for example reasonable limits of searches per user per month, or fines implemented to AI companies that are proportionate to the amount of energy and water they consume, above a certain threshold.
• Data centres should be required to make all reasonable efforts towards mitigating their environmental impact and benefiting their local communities (such as using dirty water rather than clean).
• Citizens should be widely publicly notified of the potential negative effects if a large data centre will be implemented within a specified radius of their residences, and should be able to vote on whether a data centre is implemented nearby.
• AI use should be prohibited from use in schools and tertiary educational institutions.
• Companies above a certain size, as well as the public sector, should provide a set minimum number of internship-type jobs.
• Ban AI algorithms used to review job applications and hire candidates.

This petition is not limited to the suggested regulations above.

Image credit: https://planetdetroit.org/2024/12/michigan-senate-data-center-tax-breaks/