Ban the Use of Generative AI in the State of Minnesota

 

With the introduction of AI, we are faced with a new challenge. As AI is such a radical new concept, there exist very

few laws to regulate it. As AI raises many ethical and practical concerns, this bill aims to take the safe route and

place a temporary blanket ban on all generative AI until proper legislature is enacted to regulate it. As AI is usually

controlled by large corporations, it will not be difficult to identify and block their respective models. AI companies

will be told to cease their operations within the state of MInnesota until further notice.

 

AI raises many concerns in a wide number of areas:

AI has a strongly negative effect on the environment. A single Chat-GPT request can use 10x the amount of electricity a

google search does. As a result of this, AI has a significant carbon footprint. In addition, AI requires vast amounts of

water to cool its systems. Just training GPT-3 is estimated to have consumed 700,000 liters of water. Compounding the

problem of water usage is the fact that 1 in 5 AI data centers in the US draw water from vulnerable drought stricken

areas. The massive water consumption of these data centers puts the entire region and the people living in it in serious

danger of water shortages.

AI is utilized in cybercrime, where its ability to make realistic looking photos and videos is utilized by scammers to

trick people. Many phishing scams involve AI. AI can also be easily used to spread disinformation and fake news for the

same reason. This allows people with malicious intent to trick people who don't realize content is AI into sending money

or support.

AI can reflect underlying cultural biases present in its data. As AI does not create unique content, and only pieces

together things from the internet, it can learn negative behaviors or stereotypes against certain groups of people based

on gender, race, ethnicity, etc. These biases can then be spread onto the people who read them, or other AI's who scrape

the existing AI responses for data, causing a feedback loop.

AI is a detriment to academic quality and creativity. AI is frequently used for academic dishonesty, with some models

being specifically marketed to children as cheating tools. While grammatically and technically correct, AI is unable to

generate anything more than surface level ideas, causing harm to the learning of students who decide to use it for

anything other than summary. The use of AI in scholarly papers is becoming increasingly more of a concern, as other AI

models use these same types of papers for training. This leads to a feedback loop where AI increasingly uses certain

words and grammatical phrases more, hurting the diversity of language as a whole.

AI art steals from thousands of online creators, most of the time without their permission. Generative AI will often

lift certain traits or stylistic aspects of original drawings, as it scrapes the entire internet for data. Artists used

to train AI models are almost never given credit for their work.

Deepfakes are frequently used to create harmful images and videos of people without their consent. This includes illegal

content, often sexual in nature.

AI slop is a new term being used to describe low quality content produced by AI. The presence of mass produced slop

tampers with the ability to find actually reputable information. It hurts actual people who are trying to make high

quality media and are being hidden behind AI generated material.

 

 

Artificial Intelligence (AI): Computer programs that can perform tasks usually associated with the human mind.

Generative AI: Subset of AI that specifically deals with the generation of media, such as images, text or videos.

Deepfake: Images, videos, or audio that have been altered or created using AI. They often depict real figures or events.

 

No funding will be required.

 

There will be a $100,000 dollar fine for companies that refuse to comply.

 

This bill will go into effect June 1st 2026.