IBM’s Watson AI has just gotten itself an upgrade as the company has given it an improved Tone Analyzer.
The latest beta API of Watson’s Tone Analyzer will allow the AI to identify five different emotional sub-tones: joy, sadness, anger, fear, and disgust; and upgrade over the previous versions negative, cheerful and anger categories.
Aside from emotion tones, the new Tone Analyzer will also allow Watson to pick up on social tones, or in other words, people’s personalities. Like the emotion tones, the social tones will pick up on openness, conscientiousness, extraversion, agreeableness, and emotional range.
Last but not least is the language tone. Previously known as writing tone, language tone provides feedback on how analytical, confident and tentative one’s writing is. Unlike both social tone and emotion tone, the model for language tone remains largely unchanged.
Besides identifying emotion tones, IBM has also upgraded Watson to be better at expressing emotion as well. For example, Watson will no longe reply in an upbeat tone should it detect that your writing leans heavily on anger.
Watson’s latest beta API is available for English text input. More documentation can be accessed here.
