Artificial Intelligence Removes Human Bias for Detecting Ghosts

We’re nearing the scariest day of the year—the day when devious teenagers smash your decorations, you end up with a otherwordly rum punch…

We’re nearing the scariest day of the year—the day when devious teenagers smash your decorations, you end up with a otherwordly rum punch hangover, and, some people believe, ghosts enter our plane. Yes, we’re talking about Halloween. Technology may not be able to protect your decorations or prevent your hangover just yet, but just maybe it can help detect some ghastly ghouls.

But, detecting those paranormal entities has always involved in human bias, and therefore error. How do we know if EVP (electronic voice phenomena) readings are real, or just our imagination? Redpepper is taking on the burden of settling the debate by putting the detection in the hands of an AI.

DeepWhisper, which will be running in Redpepper’s Butchertown location in Nashville from October 27th through the 31st, will use machine learning to decipher voices in the white noise picked up by a microphone. The system will run through Google’s Speech Neural Network, which is capable of understanding over 110 languages, and then parse the results. If a discernible voice is detected, it will be translated into text and placed into an SQLite data log for all the world to see.

Of course, if this does happen, there could be explanations other than chatty aparitions. Ambient noise could be picked up from the surrounding area, or various radio signals could cause interference. That’s why you, dear reader, can become a ghost hunter yourself. Subscribe to Redpepper’s YouTube channel, and you’ll be notified when a spooky specter finally speaks.

Cameron Coward
Writer for Hackster News. Proud husband and dog dad. Maker and serial hobbyist.
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles