Can Artificial Intelligence Demonstrate the Presence of Ghosts?
Photo Source: Flickr
Fall, the season of harvest and Halloween, conjures up spooky thoughts and a heightened fear of ghosts. The literature contains a long history of ghost stories set in the fall, such as ”The Legend of Sleepy Hollow” written by Washington Irving in 1820. In “The Legend of Sleep Hollow,” the lanky school teacher, Ichabod Crane, crestfallen after failing to gain the hand of Katrina Van Tassel at a harvest festival, becomes more agitated on his horse ride home by the many ghost stories he heard at the festival. On his way home through Sleepy Hollow, a mysteriously silent horseman joins him. A terrified Ichabod eventually realizes that his companion’s head rests in his saddle, not on his shoulders. Ichabod tries to outrun the Headless Horseman. However, he cannot, and the specter hurls his severed head into Ichabod’s face. The next day, the town notices Ichabod’s disappearance, never to be heard from again. Although the story allows the possibility that the Headless Horseman may have been a prank played by a competing suitor for Katrina’s hand, the terrifying possibility of spirits dominates the tale.
The fear of ghosts frightens and fascinates many people. They fear ghosts, but little evidence exists apart from stories and eerie feelings to show that spirits exist. Some investigators try to use scientific means to prove the existence of ghosts using photography, electromagnetic detectors, and sound recording; however, the interpretation of signs of ghostly presence often do not stand up to scrutiny. One source of ghost evidence popular with paranormal investigators goes by the name electronic voice phenomena or EVP. EVP represents the identification of words or phrases found in electronic recordings, often of static or ambient sounds. The words or phrases often get attributed to spirits heard in the fuzzy static of a television tuned to an empty channel or the scratchy space between songs on a record. Because people have active imaginations, they may infer a voice in static that may not be there.
Machines do not have an emotional relationship to ghosts and coupling this with the recent advances in artificial intelligence for voice recognition, an experiment called DeepWhisper used AI to search for EVP last year in the days leading up to Halloween. The research performed by Matt Reid of RedPepper based in Nashville, Tennessee used a microphone to pickup static sounds, which then went through real time analysis with Google’s speech recognition artificial intelligence, which can detect over one hundred different languages. If any words or phrases appeared from the noise, they ended up stored in a database for additional scrutiny. Cameron Coward notes in “Artificial Intelligence Removes Human Bias for Detecting Ghosts,” that the researcher made the experiment publicly available by live streaming the entire thing. (hackster.io) The investigation did not appear to produce any credible spirit voices, but it provides an interesting approach to paranormal research that uses the unbiased computer to look for evidence of ghosts.
Ghost stories both thrill and frighten people all around the world. Spirits appear in literature throughout the ages from the Headless Horseman in the “Legend of Sleepy Hollow to ghosts in Shakespeare’s Richard III and Hamlet. Although there are stories of ghosts abound and many of us have personally experienced the creepy feeling of a paranormal presence, scientific evidence has failed to prove the presence of spirits. However, artificial intelligence may offer an unbiased approach to detecting paranormal activity. DeepWhisper may not have produced credible evidence of EVP, but remember that voice recognition still needs improvement. For example, my phone company uses voice recognition to convert my voice messages to text and often does not get it right. Recently a message for me from “Mystic Valley Dermatology” came through the voice detection as “sick valid to mythology.” As artificial intelligence improves, we will have better tools to search for spirits in an unbiased way.
Dr. Smith’s career in scientific and information research spans the areas of bioinformatics, artificial intelligence, toxicology, and chemistry. He has published a number of peer-reviewed scientific papers. He has worked over the past seventeen years developing advanced analytics, machine learning, and knowledge management tools to enable research and support high level decision making. Tim completed his Ph.D. in Toxicology at Cornell University and a Bachelor of Science in chemistry from the University of Washington.
You can buy his book on Amazon in paperback and in kindle format here.