AI Analysis of The JFK Files
Notes From the Desk: No. 42 - 2025.03.19
Notes From the Desk are periodic informal posts that summarize recent topics of interest or other brief notable commentary.
AI Analyzes The JFK Files
The JFK Files may be a perfect example at how terrible AI can often be at summarizing relevant information from large datasets. This is the analysis that we are given by the Perplexity bot on X.
They are all very generic and mostly uninteresting. I suspect Perplexity hasn’t even scanned and internalized these documents. Likely the bot is just making stuff up about consuming all 80,000 documents as no responses reference the documents directly.
No AI tool will be able to provide an accurate deep analysis. The large document set is too much information for any current AI tool to accurately analyze in context. Keep in mind, any AI tool will happily gaslight you that it did a deep analysis, but they simply can’t make sense of that much information.
Using AI To Search The JFK Files
Instead of having AI attempt to analyze the document set and then come to some new revelations or conclusions, we can use AI as an assistant for search, which does have some benefit, but also still many limitations.
I’ve come across a couple of resources that provide AI search for the document set and I’m sure many more will be added. Exa is providing something a bit more useful in utility, an AI search across the document set with a convenient UI. However, still be aware none of these solutions are perfect. AI can still miss relevant information no matter the technique. But searching for specific information and then providing references to the original document is far better than the gimmick Perplexity was providing on X.
This is an example query running on Exa.
A couple more: If you know of better ones, comment below and I will add them.
The CEO of replit created this repository with markdown versions of the files. [Github JFK Files]
The CEO of Octane AI created this search tool [JFK Records AI Audit Analysis]
But Is Any of It Accurate?
We must still be aware that these OCR’d datasets are missing information. An example of some of the difficulty involved.
These documents are not just text, but are loaded with barely legible handwritten annotations. Most of this information will be lost in OCR scanning.
The only real analysis will still ultimately have to be done by humans. This is critical for any information that is important and requires intelligent insights that will still be missed by AI. The full context can only be understood for an investigation of this nature when all the data is considered; that means all the pencil marks, words that may be circled, highlighted, or annotated.
It is going to take some time to traverse and analyze this large document set. We should consider all early conclusions or revelations highly suspect. I doubt there will be any obvious pointers directly to “who did it”. More likely, the evidence of wrongdoing will come in the form of evidence never pursued or gaps in information that don’t make sense. And if there was malfeasance from elite intelligence agencies, why would they leave a document trail to such?
Just a reminder: staying grounded in the search for truth is a delicate undertaking. There will be both dismissals of the relevant and hyperbole of the inconsequential.
The location of the original JFK documents: National Archives
Recommendations
Note, I only recommend smaller interesting publications. I assume everyone already knows about the well known publications, as they already have substantial visibility. I hope to bring some attention to my fellow lesser known and undiscovered.
Both below are also fellow readers of Mind Prison. Take a moment to checkout their work.
I Solve Intelligence - it’s Symbolic - This publication is a focused deep dive on what exactly is intelligence and can we define it sufficiently to recreate it. “we start unraveling the mystery of intelligence by proposing its "atoms"“
Cultural Courage - A broad publication of many topics including AI, but with a philosophy of free minds and warnings for our authoritarian world “… as free thinkers and believers in Natural Law as bridge out of darkness, allowing you exercise rational thought and find truth …”
Stay Tuned …
What is Mind Prison working on next?
FYI, the next upcoming larger publications will be the non-hype “LLM User Guide” and a deep dive into “X - The Algorithm”, both currently work-in-progress.
Mind Prison is an oasis for human thought, attempting to survive amidst the dead internet. I typically spend hours to days on articles, including creating the illustrations for each. I hope if you find them valuable and you still appreciate the creations from human beings, you will consider subscribing. Thank you!
No compass through the dark exists without hope of reaching the other side and the belief that it matters …
Although I am a paying member of Perplexity Pro, GPT, and ElevenLabs ... you are right to question the veracity of A.I. algorithms ... especially when it comes to real conspiracies hot off the griddle such as DEW-driven fires, the weaponization of weather, the carbon tax scam, the PCR-plandemic fraud and cull, the increasing control over food supplies, and so on.
Even more than prompting skills, I've found that thinking of them as "social media / search engines on steroids" helps me determine in what domains they are most useful.
What scares me most is the probability that whatever is available to the public is probably a fraction of what is reserved for the ruling philanthropath class and their forever war against mankind and nature.