Artificial intelligence has changed the way we can work with information. At The Ace Holes, we use AI as an editorial tool when researching, writing, editing, and updating articles. It allows us to process larger amounts of information and work more efficiently, but it does not change our responsibility for what we publish.
The most important distinction in our approach is between AI as a tool and sources as documentation. An answer from an AI model is not a source in itself. If AI provides a specific piece of information that we want to present as fact, we need to determine where that information comes from and evaluate the original source behind it.
From idea to research
AI can become part of the process before we even begin writing. Once we decide to cover a particular topic, we can use it to map out the available information, identify questions the article should answer, and find areas that require further investigation.
This is particularly useful when a topic involves several products, manufacturers, rules, statistics, or other information spread across multiple sources.
AI might, for example, help us discover that two sources provide different specifications for the same disc. What matters at that point is not which answer AI chooses, but the fact that there is a discrepancy that needs to be investigated.
Used this way, AI can help make our research broader and more systematic. The facts we ultimately use still need to be traceable to their sources.
We look for the original source
When we find information we want to use, we try to trace it back to the original source whenever practical.
If we are writing about a new disc, the manufacturer will normally be the natural source for information about its release, available plastics, stated flight numbers, and technical specifications. If we are writing about rules, the actual rulebook or the organization responsible for those rules carries more weight than an article that simply repeats them. Likewise, when we use statistics, we try to find the organization or database that originally published the numbers.
That does not mean an original source automatically provides every answer.
A manufacturer can document the flight numbers it has assigned to a disc, but that is not the same as documenting exactly how the disc will fly for every player. A manufacturer may describe a disc as very overstable, while players may experience it differently. In those situations, the article needs to distinguish between information provided by the manufacturer and observations or opinions from other sources.
This is standard source evaluation, but AI makes the distinction particularly important because it can combine information from different contexts into a single answer that may appear more definitive than the underlying sources actually are.
We verify claims, not just the writing
One of the challenges with generative AI is that accurate and inaccurate information can be presented with the same level of confidence. A well-written sentence is not necessarily a correct sentence.
Our fact-checking therefore focuses on the claims within the article.
If an article states that a disc was released in a particular year, we should be able to verify that year. If we provide a specific diameter, weight limit, plastic type, or other technical specification, that information should also be traceable to a relevant source. The same principle applies when we write about tournaments, rules, companies, people, or historical events.
This follows a well-established journalistic principle: information should be verified before publication, and original sources should be used whenever possible. Those principles are included in the Code of Ethics.
takes a similar approach specifically to generative AI, treating information produced by AI as unverified material that remains subject to normal editorial judgment and sourcing standards.
That reflects how we approach the technology as well: AI can point us toward something worth investigating, but AI itself is not proof that the information is correct.
What happens when sources disagree?
Checking sources does not always lead to a single, definitive answer.
This is especially true when writing about disc golf equipment. Two discs with identical flight numbers will not necessarily behave the same way, and different players can have very different experiences with the same model. Plastic, weight, manufacturing variation, wear, throwing technique, and arm speed can all affect how a disc flies.
In situations like these, it is important to distinguish between documented information and subjective observations.
We can establish which flight numbers a manufacturer assigns to a disc. We can describe how the manufacturer positions that disc. We can also use player experiences and other sources to provide a broader picture. But those experiences should not automatically be presented as measurable facts.
When sources provide conflicting factual information, we try to understand why. One source may be outdated, the product may exist in several versions, or the sources may simply be describing different things. If we cannot resolve the discrepancy with reasonable confidence, it is more accurate to acknowledge that uncertainty than to let AI choose one answer for us.
AI as a writing and editing tool
Once the research is done, AI can also assist with the writing process.
We use it to help organize information, create a logical structure, improve clarity and flow, reduce repetition, and make complicated subjects easier to understand. AI can also assist when translating and adapting articles between languages.
This is one of the areas where the technology can save considerable time without replacing the editorial process itself.
Major news organizations also use AI for some of these purposes, for example, allows AI assistance for tasks including early-stage research, document summarization, grammar, search optimization, and headline suggestions, while requiring AI-generated output to be reviewed and edited before publication, also states that generative AI is used as one of several tools in reporting, writing, editing, production, and publishing.
AI can therefore be involved in several stages of creating an article. That does not mean a polished piece of AI-generated writing is automatically ready to publish.
AI does not make a source more reliable
Another important principle is that AI does not improve the quality of the information it is working with.
If the original source is poor, outdated, or misleading, AI may be able to present that information more clearly, but the underlying problem remains. This is why we evaluate a source independently of how we found it.
We look at who published the information, when it was published or updated, where the information originally came from, and whether the source has interests that are relevant to the subject.
This does not mean commercial sources should be dismissed. A manufacturer is often the best source for its own product specifications. A retailer may be a useful source for current availability or products being sold. What matters is whether the source is being used to support something it is actually in a position to document.
Source evaluation is therefore about more than determining whether a source is trustworthy. It is also about understanding what that particular source can reliably tell us.
AI-generated images are illustrations
We also use AI to create visual content for The Ace Holes. Here, the distinction between illustration and documentation is important.
An AI-generated image can illustrate a disc in flight, a general situation on a disc golf course, or a concept that would otherwise be difficult to photograph. But that image does not document an event that actually took place.
When the product itself is important to the story, a real product photo serves a different purpose than an AI-generated illustration. The same applies to photographs of actual players, tournaments, courses, and events.
This distinction is becoming increasingly important as AI-generated images become harder to distinguish from photographs, for example, maintains specific standards separating AI-generated illustrations from journalistic photography and does not allow generative AI to add or remove elements from news photographs.
The editorial responsibility remains ours
AI can be involved in many parts of the work behind an article. It can assist with research, compare information, help organize a story, improve language, translate content, and create illustrations.
What it cannot do is take responsibility for what we publish.
The Ace Holes chooses the topics we cover, evaluates the sources, decides what information is relevant, and publishes the finished article. If a fact is wrong, it does not become any less wrong because it was originally suggested by an AI model.
The goal is therefore not to remove people from the process. It is to use the technology where it is useful, giving us more time to investigate, verify, and improve the content we publish.
Our approach can be summed up simply:
AI helps us find, compare, organize, and communicate information. Sources provide the evidence. The Ace Holes remains responsible for what we publish.