Voice conversation analysis method and apparatus using artificial intelligence

An electronic apparatus for analyzing a voice by using an artificial intelligence learning model, the electronic apparatus comprising:

• a communicator;
• a voice receiver;
• a voice main part obtainer;
• and
• a processor,
• wherein the processor is configured to:
• control the voice receiver to receive input of voice,
• control the voice main part obtainer to obtain information on the received voice,
• identify whether a voice model data pattern matching the information on the received voice is present in a database comprising a plurality of voice model data patterns corresponding to a plurality of categories, wherein the voice model data pattern comprises a mechanical sound determination model, a gender classification model, an age classification model, and a language classification model,
• based on the voice model data pattern matching the information on the voice being present in the database, identify category information on the received voice as category information corresponding to the voice model data pattern,
• determine at least one operation mode corresponding to the category information based on at least one of obtained output information by inputting the voice to a plurality of artificial intelligence models corresponding to a plurality of operation modes or setting information obtained by a user along with the identified category information,
• based on the voice model data pattern matching the information on the voice not being present in the database, identify the category information as an uncategorized category and determine the operation mode to be a pre-set basic mode,
• based on the determined operation mode being a voice phishing prevention mode, obtain content, in which the voice is textualized, and a voice feature information from the voice, wherein the feature information comprises a shaking of the voice, tone, or number of phonemes per unit time,
• control the communicator to receive an output control instruction from an external server by inputting the content and the feature information to the artificial intelligence model corresponding to the voice phishing prevention mode,
• perform at least one function of the voice phishing prevention mode based on the control instruction, wherein a function voice phishing prevention mode comprises displaying a report button, limiting financial applications, or a voice phishing alarm,
• based on the basic mode being an automatic mode, input the voice to the plurality of artificial intelligence models corresponding to the plurality of operation modes and determine an operation mode corresponding to an output information based on the output information, and
• determine a category information corresponding to the operation mode, and update a voice model data pattern corresponding to the category information based on information on the voice.