Presidential Bill:1/3: Difference between revisions
From The Commonwealth of Acropolis
More actions
No edit summary |
No edit summary Tags: Manual revert 2017 source edit |
||
| (3 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
{{Act template|title=The Generation Act}} | |||
== Title I - Findings == | == Title I - Findings == | ||
Latest revision as of 20:34, 15 September 2026
Be it enacted by the Senate of the Commonwealth assembled and the President of the Commonwealth presiding.
This act shall be titled as "The Generation Act."
Title I - Findings
Sec. 1. Naming
- It has been found that, in every case, so-called "Artificial Intelligence" is not to be classed as intelligence. Therefore, the Commonwealth recommends it be called a "Large Language Model", or if such a term as "AI" is to be used, the "I" shall be placed in quotation marks.
Sec. 2. Immediate Detriment
- It has been found that students use Large Language Models in many cases to complete assignments, due to the difficulty presented with such work, the inability for present teachers to teach in a way their pupils comprehend, and the fact that work is assigned, typically, for total accuracy rather than for understanding. In doing so, they are worsening their own abilities to understand and complete work on their own for that assignment.
Sec. 3. Factual Misdirection
- It has been found that LLMs and other Generative Algorithms frequently provide information that is false, opinionated, or misleading.
Sec. 4. Media Generation
- Media generated by diffusion models, including text-to-image (TTI), text-to-video (TTV), and text-to-audio (TTA) models, are often posited to "replace artists." However, media generated by such models tend to, in every case, lie in a sort of "uncanny valley" of media. It is also becoming increasingly simple to spot generated media, in addition to it being at a significantly lesser quality to that of real, bona fide, genuine human works.
Sec. 5. Copyright and License Violations
- It has been found that, in many cases (except for Public Domain or where permitted), data used to train Large Language and Diffusion models is obtained or used in a way that is illegal. For example, if a Large Language Model that specializes in code is trained on a GPL licensed project, yet does not attribute that work under GPL, any code generated as a result of that code is deemed a license violation.
- Additionally, if a work is licensed under CC BY-NC-ND 4.0, and that work is used to train a Large Language Model, one could view any subsequent created work as a derivation, meaning it is too a violation of the license, due to the malattribution of work, use of the work to further a monetary goal, and making a new version of that work.
Sec. 6. Data Centers
- It has been found that Data Centers, the locations by which Large Language Models and diffusion models, utilize a large number of natural resources, as well as contributing a large number of pollutants and general detrimental effects to the environment, including humans.
- One such natural resource is water. Water, due to it being a readily available thin liquid with a higher thermal conductivity than that of air, is exceedingly good at removing waste heat. However, due to the amount of energy it requires to cool the water therein, it is often easier to simply dump the water such that the city handles it. Yet, this method removes water from local watersheds and habitats, thereby damaging the ecosystem.
- Another such resource is that of land. Land, by the very nature of itself, is finite. Due to data centres taking up space congruent with that of National Parks, they too usurp land from both the people and the wildlife therein.
- Such pollutants produced include that of hydrocarbons and Greenhouse gases. These pollutants cause harm to the environment, the ecosystem, and overall that of the planet.
Sec. 7. Summation
- In summation, in accordance with data presented above, it has been found that, Large Language Models, as well as other Generative Models, tend to grant more harm than good to humanity.
Title II - Crimes
Sec. 1. Redress of Findings
- In accordance with the findings aforementioned, it shall be noted that Large Language Models are to be considered harmful to many degrees, and shall thereby not be utilized.
Sec. 2. Crime of Usage
- Due to the severe detriment caused by the use of Large Language Models and other Generative Algorithms, it shall be regarded that use of such systems be considered illegal under all applicable Commonwealth Law, and never regarded as used in "positive spirit" or "well intentions."
- The punishment for use of such systems shall be determined by the severity of use, as well as other factors including emotional damage caused by those who had to witness such use. The maximum fine per violation shall be no more than that of 50 United States Dollars granted to a fund provided under this act.
Title III - Fund
Sec. 1. Creation of Fund
- It shall be that the Clerk or President of the Executive hold and collect funds in accordance with any violations of laws provided in this act.
Sec. 2. Regulation of Funds
- The purpose of the monies collected may be appropriated by the Senate. The ledger of the monies collected may be regulated by the Clerk or President of the Executive.