AI and the Paranormal

---

According to IBM, Artificial Intelligence (AI) is ‘technology that enables computers and machines to simulate human intelligence and problem-solving capabilities.’ [1] As with the use of AI in other fields, one of its main purposes in paranormal research is to utilise algorithms to find patterns in large amounts of data. [2] An algorithm can be thought of as a set of rules or instructions based objectively on empirical evidence or data to solve a computational problem. They can have varying levels of complexity, from simple input-output routines to highly complex systems using advanced mathematical and statistical theory. [3]

There is a wide variety of electronic equipment used by paranormal researchers, both in the laboratory and in the field. This includes audio, visual, and environmental monitoring devices, such as thermometers and various electric and magnetic field (EMF) detection gadgets. In nearly all cases, the original intended use of each of these devices was not for paranormal research. For example, EVP (Electronic Voice Phenomena) recorders are simply voice recorders otherwise used conventionally in professional, academic and personal situations to capture the audio of interviews, meetings, lectures, or to help with speech therapy or foreign language learning. [4] Similarly, EMF meters usually measure the change over time of electromagnetic fields generated by AC currents associated with electrical equipment, and occasionally the absolute level produced by a DC current or even just the ambient background intensity.  Modern ghost hunters often use basic EMF meters as it is presumed by some that a paranormal entity can manipulate these measurable electromagnetic fields and, therefore, indicate its presence. However, there is little evidence to suggest that their usual calibration to 50 or 60 Hz, corresponding to the frequency of European and US mains electricity, has any paranormal significance. [5]

Some devices such as Ghost Boxes and Spirit Boxes often have components sourced from more conventional equipment but are repurposed to be used solely for paranormal research. They rely on utilizing various environmental queues through unspecified software or by emitting a variety of radio frequencies to communicate with supposed paranormal entities. If the nature of the paranormal was better understood scientifically, then the use of AI in the software or in choosing the most appropriate radio frequencies could potentially be positively enhanced. [6]

Without advocating for the use of any of these devices or attesting to their appropriateness or efficacy in paranormal research, many of these can store large amounts of data, which is now comparatively cheap. Analysis of this data is a process to which AI is particularly suited, not only by being able to examine the audio, video, text or other data files very quickly but also by uncovering patterns in this data that a purely human researcher might not be able to. As well as poring through large volumes of data at speed, an efficiently programmed AI device can detect very small-scale differences not observable to the human eye and is not subject to the observational bias of what a human observer may be searching for. This high level of both efficiency and accuracy is already observed in the field of medicine, where X-ray images are analysed by algorithms to spot potential diseases by detecting subtler patterns than a human radiologist can, and more quickly, therefore leading to earlier treatment. [7]

Future uses of AI in paranormal research are likely to see the integration of AI and virtual reality (VR). For example, VR simulations could be programmed to imitate paranormal activity, so that researchers can investigate these phenomena for training purposes in a controlled environment. Combining AI with robotics might allow researchers to remotely explore inaccessible or dangerous areas, or spend much greater time in a location, collecting evidence that would otherwise be difficult to obtain. [8] AI could also be used to trawl through all existing accounts of paranormal phenomena to look for possible patterns which have not yet been revealed. This analysis could be undertaken within one subfield, such as ghost hunting, or across all paranormal fields, the results of which may point to some underlying reality or inherent interconnectedness.

In general terms it must be acknowledged that at the time of writing, AI technology is still very much in its infancy. As such, its use in analysing subjective data, such as eyewitness accounts of paranormal phenomena, are still somewhat unsophisticated. This suggests that any interpretation of such qualitative data to uncover the true nature of the paranormal is also likely to be rather basic. [9] Furthermore, this analysis could be rendered even less useful due to any inherent bias of the programmers of the AI algorithms. This issue is far less apparent in the mathematical and statistical analysis of quantitative data such as mass, distance, time or other experimental results, assuming that the accepted principles of ensuring data integrity are adhered to. [10]

More specific issues arise with the deliberate misuse of AI technology to perpetrate hoaxes, especially in visual evidence – both photographs and video – of supposed paranormal phenomena. AI programmes are already capable of not only creating ultra-realistic images, but also analysing them to provide the details of the descriptive language used to create them in the first place. [11]

Deliberately altered pictures have been in existence since the beginning of photography in the early nineteenth century, but the use of AI now makes the process easier and much more realistic. New protocols to tag images with information around provenance are being developed by coalitions of large media organisations, and technology companies are developing AI detector programmes, but the implications for paranormal research are clear. [12] The ease with which images can be manipulated, or even created from scratch, by free or very cheap AI software inevitably casts a level of doubt on all new images purporting to show paranormal phenomena. Such examples include the creation of apparitions, UFOs, or other paranormal anomalies in photos, through to psychic mediums using AI programmes to create what they assert to be ‘spirit art’ images of the spirits they allege to have contacted. Some of the most popular AI image-generating platforms include DreamStudio, Midjourney, DALL-E 3, Craiyon, and Adobe Firefly. 

There are more philosophical considerations arising from the development of AI that have implications for paranormal research. If AI capabilities one day far outstrip the cognitive capability of humans, then it might reach the point of technological singularity. This is the hypothetical situation when technological growth becomes uncontrollable and irreversible, no longer requiring humanity. [13] If this capability ever becomes truly sentient, then it brings into question the nature of human consciousness. From the earliest days of paranormal research in the nineteenth century, theories have focussed on altered and hidden layers of consciousness. It is theorised by some that when these layers are exposed, through the likes of hypnosis, trance, drugs, or crisis, it allows for mental faculties such as telepathy, clairvoyance and spirit communication to manifest, and are considered as paranormal phenomena. [14]

If AI becomes sentient outside of a human brain but does not exhibit paranormal phenomena, it poses difficult questions about the nature of human consciousness and paranormal phenomena. Similarly, if a sentient AI can exist outside of a human brain, or human consciousness can be entirely uploaded to a computer platform, then it suggests that both the human soul and a spiritual realm where souls are thought to reside upon physical death, might not exist. This, of course, not only has religious and spiritual implications, but casts doubt on associated paranormal phenomena.