PRAVDALIST.AI

▓▒░ expose deception
Оберить мову | Select Language | Выбирете язык

Limitations

  • Ability to work only with the following languages:
    • English
    • Ukrainian
    • Russian
    • with any other language in the future
  • Processing time is approximately 2 times longer than the length of the audio recording
  • The recording should be relatively clean, without significant noise
  • The speaker's speech should not have any clearly expressed multiple defects
  • Interlocutors, when there is more than one, should not speak at the same time
  • The system is capable of detecting any deception, even if it is:

  • **Brockaru Trap** (when a speaker’s manner of speaking consistently triggers deception detection, making it impossible to determine when they are actually lying). To address this issue, the speaker must undergo preparation as follows: Ask questions that trigger a positive psychological response from the speaker (the person being interrogated/interviewed), such as: "What is your mother's name?", "What is your name?", "When is your birthday?", or "What is your favorite dish?" You could also ask about the place where they spent their childhood, and similar questions. If the answers to these questions are detected as truthful, then there is no Brockaru Trap, and any further deception can be considered genuine. This method should be used when nearly all answers appear deceptive, raising suspicion of a Brockaru Trap. However, this still does not eliminate the effect of detecting a pervasive deception.
  • False deception or True lies (speech that is defined as a lie, but has no motive for deception and is not a lie in itself - a new phenomenon discovered in the process of developing this technology for detecting deception)
  • Lies in a “TV show” (this is a lie without motivation, but conditioned by the circumstances in which the speaker finds himself)