Generative AI and AI has truly changed the world. But how useful is it in the paranormal? Is using generative AI in spirit communication even possible? Let’s take a deep dive and find out.
In my day job, I work as a learning experience designer and I work for one of the leading cloud computing companies. This means I have extensive familiarity with the various AI platforms, companies, and tools in the market. My experience gives me unique insight into how generative AI operates, as well as its benefits and limitations.
My technical knowledge, combined with my experience in the paranormal field, informs my perspective on why these tools should be approached with caution when applied beyond data review and investigation. While these AI systems can offer amazing capabilities for processing and analyzing large amounts of data, using them in contexts like spirit communication raises important concerns.
How AI Functions
When it comes to using generative AI for spirit communication, several issues arise, particularly because of how AI functions. One of the primary issues is that AI models are trained on vast datasets of human-created content, meaning they inherently reflect biases.
Since we have yet to establish a baseline definition of what a spirit is, how it communicates, or what constitutes a genuine spirit voice, it becomes nearly impossible to train AI to accurately identify or interact with spirits. AI needs something tangible to create an accurate output, and without a concrete framework, it’s left to generate responses based on human patterns and expectations, which may not reflect paranormal reality. This can skew results and lead investigators to believe they are receiving meaningful responses when, in fact, they are not.
This makes it likely that an AI will misunderstand or misrepresent spiritual communication, as it is designed to generate plausible human-like responses based on existing data rather than interpreting actual paranormal phenomena. Because of this, responses generated by AI are influenced by language patterns and expectations set by human interaction, which can skew results and lead investigators to believe they are receiving meaningful responses when, in fact, they are not.
The Problem of AI Outputs
Another significant issue is the ambiguity and interpretability of AI outputs. Generative AI lacks the ability to discern between authentic spiritual communication and random noise, human input, or algorithmic content. It may produce text that sounds coherent and convincing, but it is simply an automated response created from patterns. This opens up the potential for investigators to be misled or to attribute significance to something purely generated by the AI without any connection to the paranormal. Since AI can fabricate content, distinguishing between genuine spiritual experiences and machine-generated material becomes increasingly difficult.
Ethical Concerns
Ethical concerns also arise when using AI in spirit communication. There’s a risk that the process could trivialize or disrespect both spirits and investigators. AI-generated responses could create a false sense of communication. This may be harmful, especially to those seeking closure, comfort, or understanding through spiritual means. Relying on AI in these cases can lead to unintended consequences. This includes deep emotional harm for those who believe they are receiving messages from the other side.
Moreover, using AI for spirit communication could lead to a loss of the human element. The human element is often central to the investigation process. Paranormal investigations are subjective experiences that rely on human intuition, empathy, and personal connection. AI lacks the emotional and intuitive capacity to engage meaningfully with spirits. This means it cannot replicate the personal interactions that are often key to understanding or communicating with the paranormal.
Stolen Content
On top of these concerns, the ethical implications of AI and Generative AI extend beyond spirit communication. The technology itself raises serious issues. This is because it is built on the wholesale theft of material—writing, art, and more—in blatant violation of copyright law. The datasets used to train these AI models often include copyrighted works without permission, which not only harms creators but also degrades the credibility of the outputs. The reliance on stolen content compromises the legitimacy of any conclusions drawn from AI-generated material, effectively poisoning the well for anything that could be considered remotely credible. Additionally, the technology is environmentally problematic due to the high energy consumption required for training large AI models, contributing to carbon emissions and other environmental impacts.
Risk of Deception
Finally, AI’s ability to generate text that seems plausible increases the risk of deception. This could lead individuals to believe they are engaging with spirits when they are merely interacting with a machine. This risk of false conclusions makes using AI in spirit communication questionable and potentially damaging to the integrity of paranormal research.
How Generative AI Could Be Helpful
Generative AI could still play a valuable role in paranormal investigations, particularly when reviewing evidence after an investigation. AI can help analyze large amounts of audio, video, or photographic data more efficiently than humans. It can spot patterns or anomalies that may otherwise go unnoticed. Also, it can be trained to flag potential EVP recordings or visual distortions for further investigation. This means streamlining the review process and making it easier for investigators to focus on key findings. In this way, AI serves as a useful tool for enhancing evidence review. This allows human investigators to dig deeper into their findings with a more focused lens. However, this use case is distinct from attempting to communicate directly with spirits. As of now, the limitations and ethical concerns remain significant.
Very interesting and thought provoking. As usual, thanks for sharing.
Regarding the use of AI in direct communication I agree completely. The baseline used by the system is subject to bias based on the source of the data used to create it. Thus any output is going to be subject to error introduced by that bias. I would take it a step further in using it in analysis as well. You mention flagging points of interest. On what would we base the criteria for flagging? That in itself introduces the potential to miss something if the data used was not based on real world sampling. And what do we have as a basis? Consider EVP for example. We have no 100% confirmed EVP samples (Although we do have some that might be 99%, that in itself leaves a 1% chance of error.) This opens a possibility for either missing a valid input or also a false positive. This is no better than what any investigator can do using current methods being employed. There are enough problems standing in the way of serious paranormal investigation without introducing the ones created by another potential source of error.
As a paranormal investigator from the early 2000’s when “spirit boxes” were not the norm, I have always misstrusted using them as I felt I never could be sure of who was coming through. Now reading this makes sense as we really cannot be sure whom or what we are really communicating with. With AI controlling the content, I am more skeptical of the content. That only leaves audible communication which is heard in the moment and the digital recorder, my now only tools. If I could do automatic writing or was a psychic that would be the second choice.