Office of the Bravo Fleet Judge Advocate General

The Defendant, Vausees Vax, is charged with violating the following Article(s) of Conduct and Reprimand as defined by Section 6 of the Bravo Fleet Judicial Code.

• 1x Plagiarism (Plea: None)

  Members must not submit any plagiarized work as their own. Work of others, including content generated by software or a website, may be used in a member’s work provided that the sections of the work that are not wholly original work of the member and the sources thereof are identified and disclosed.

• 1x Cheating and Exploitation (Plea: None)

  Members must not willfully or knowingly exploit, abuse, or otherwise tamper with official game play or official activity in the form of competitions, tests, or other activities. As used in this section, exploitation means the unsanctioned use of bugs or loopholes in gameplay for the purpose of gaining an obviously unfair advantage. Cheating means the use of any dishonest or unfair act for the purpose of gaining an advantage. Unauthorized use of the same submission in two separate competitions is a form of cheating.

On November 10th, 2023 the following story submissions by Vausees Vax (VivdShado) were flagged during Bravo Fleet Senior Staff review as containing likely AI generated content.

• https://bravofleet.com/story/93140/
• https://bravofleet.com/story/93149/
• https://bravofleet.com/story/93275/
• https://bravofleet.com/story/93562/

The staff was on high alert for any AI generated writing in this case because there had been other cases of content written by fleet members previously (including story comments) that appeared to be AI and, when put through AI detection software, was marked as probably AI content. The Bravo Fleet Creative Integrity Policy had been announced publicly several times, including a reminder news post on August 26th, 2023, given the rise of AI generated content being posted to the site. As such, these posts were marked for further investigation following the closure of competitions during the Fleet Action.

Before any additional investigation into the matter could be engaged with, on November 14th, 2023 the accused admitted to the use of ChatGPT to “help” him with his stories in open chat in the Task Force 72 Lounge on Discord. This action was in direct violation of the Judicial Code for not disclosing the use of content that was not generated by the member in question with the submission. Additionally, this violates the Creative Integrity Policy and its ban on the use of AI generated content in fiction submissions.  Additionally, the fiction submissions above were entered by Vausees Vax into the We Are the Borg: Writing Ribbon Race competition. This violates the Creative Integrity Policy prohibition on the inclusion of AI generated content in any competition entry and additionally the Judicial Code prohibition on cheating an exploitation by including content not written by the entrant.

At this point the Bravo Fleet Executive Officer was authorized by the Bravo Fleet Commanding Officer to immediately impose a 30 day activity ban from any writing of competition activities within the fleet. The defendant was notified of the charges that were being filed against them and what possible punishment was being sought by the Judge Advocate General. The defendant was provided sufficient counsel from the Bravo Fleet Defender and was made aware of their rights under the Judicial Code. Due to the nature of the charges, the possible penalties would not exceed those of a one-grade demotion, a Letter of Reprimand, one year probation, or a combination of the above. Therefore, pursuant to Judicial Code Section 3 – Process of Adjudication, a Trial by Judge was to be set.

The defendant did not respond to the charge and thus entered a plea of NOT GUILTY by default, and the case proceeded immediately to a Trial by Judge without defendant participation as directed by the Bravo Fleet Judicial Code Section 3 – Process of Adjudication, Failure to Plea.

This is the first instance in Bravo Fleet for the Creative Integrity Policy to be enforced, and as such offered a number of unique challenges. Detection software for AI content is not 100% correct in all cases, and is often best used to compare changes in writing content over time as opposed to using it on isolated writing samples, IE if written content by an author suddenly changes to flag as AI when it did not before these results can be trusted more than others.

This specific case, however, while initially setting out to define what a burden of proof is for a member to rebut charges of the use of AI or what burden the JAG has to meet to prove their case, cut through all of those concerns in the face of an unrepentant confession by the accused.

Most importantly, however, is what this case means for Bravo Fleet moving forward. Where is the line between spell check/grammar check and AI writing content? It is the opinion of this body that the dividing line must be here: there is a difference in correcting a mistake and having the “machine” improve what is otherwise serviceable content. Having an AI create an outline that you will then use to shape your own writing is acceptable whereas having your writing serve as the “outline” for the AI to build upon is not. In this case the defendant directly admitted to the use of AI to add content to his writing, a clear violation of even the most narrow interpretation of a ban on AI content.

Beyond the specifics of this case, however, the defendant did raise concerns about the method of notification during cases. The defendant was notified via email of charges being filed against them, as well as their rights under the Judicial Code as sent by the Bravo Fleet Defender. They claimed that this was unfair as they only check their email once a month, and that they should have been notified via Discord as they missed the notification of their charges in time to enter a plea. However, all JAG communication will always take place via email as it provides a permanent, immutable record of communication that Discord lacks as email does not allow for messages to be edited or deleted after the fact. Based on this feedback, however, the defendant was given notification via Discord that the verdict had been reached and that the 72 hour window to file an appeal had begun. Email should be expected to be used as an official fleet communication channel. Members are required to provide an email address to be listed on their dossier and the site itself sends a variety of email notifications to the membership (promotions, medals, etc). In the future, the JAG office may look to require notification via email and the fleet’s primary chat system. In this case, however, it was not a requirement per the Judicial Code.

In reviewing the facts of this case, it was important to examine the underlying ideas that underpin the Bravo Fleet Creative Integrity Policy. Why a prohibition against AI is important (outside of the context of a competition at least) is sometimes hard to quantify or imagine. This is not a place where any of our members are depending on their writing for their income, so why treat it like someone’s livelihood is being replaced by machine learning you might ask. The issue here is that we are a community for people to come together and share their writing with others, and share the activity of writing with others. If AI generated content is allowed, then there is no longer an incentive to produce and share content with one another. There is no ability for authors to actively improve their writing as AI can just punch up their content for them. In short, AI content poses and existential threat to our organization just as it does for those that make their living as authors. If your hand crafted writing is just going to be buried in an unending sea of impersonal, mass generated content, why even try?

Additionally, in this case the entries were also submitted as entries contained under the umbrella of a Fleet Action. These major events are the flagship of Bravo Fleet’s new (yes, even four years in) competitive initiative, and our members expect a level playing field. Article II Section 1 of the charter even states that every member has the right to be adequately acknowledged and rewarded for time, activity and efforts. When members use AI to take a shortcut to either improve the quality or word count of their writing, they are skewing the playing field to devalue the time, activity, and effort of all of the other members that are playing fair. As such, AI use in competitions is even more of a violation given that it is a direct infringement on the rights of other members.

Beyond that, in the specific facts of this case, it was apparent from the conduct and statements of the accused that the member in question saw nothing really all that wrong with what they had done. Even after they were directed to the policy, the accused still maintained that the AI didn’t generate the stories, they just used the AI to add details so they’re still the author. As JAG, I will almost never look to bring a case against a member simply because they technically violated a rule. The foundation of the JAG process is that it works first and foremost as a tool to make members understand that their conduct is wrong, to help them learn from these mistakes and become better members moving forward. A member being ignorant of a rule, but understanding what they did was improper goes a long way towards softening the stance of the JAG, Investigator, and Defender in these matters (even as ignorance is not a defense, especially given the number of times the Creative Integrity Policy was publicly announced). This, however, was not the case in this matter. As such, it was agreed upon by the full JAG Office that the following sentence would be imposed in this case:

• 1 Grade rank demotion to the rank of Commander
• General Probation of 12 months
• 1 Letter of Reprimand