Extending Five Factor model to Chatbot Personality
When it comes to human personalities, I relate to the five personality traits theory aka five factor model (FFM). I will not go too detail into this as this blog is more focused on how this model can be extended to defining/designing chatbot personality.
The Five factor model defines the taxonomy for personality traitsby using the below five factors as listed below –
- Openness to experience
– defined by the degree of curiosity, creativity and preference for novelty and variety involved. It is also described as the extent to which a person is imaginative or independent and depicts a personal preference for a variety of activities over a strict routine.
- Conscientiousness– defined by the degree of efficient and organized. This is usually reflected by the planned vs spontaneous behaviors. In other words, low conscientiousness means more flexibility and spontaneity but can also appear low reliability.
- Extraversion– defined by the degree of outgoing and energetic. This is usually reflected by more positive emotions, assertiveness and sociability.
- Agreeableness
– defined by the degree of friendly and compassionate.This is usually reflected by the measure of trust, helpful and even the level of submissiveness. Low agreeableness also seen as most often competitive/challenging and as argumentative.
- Neuroticism– defined by the degree of sensitiveness, emotional stability and impulse control. Higher neuroticism is viewed as more unstable, and insecure.
Phew…now that we have covered the traits, let’s see how we can relate these traits to chatbot personality or chatbot persona by using the dialog and intent
| Factors | Intent or Dialogs that could be used to identify chatbot personality | Example chatbot personality BOTs |
| Openness to experience | This trait could be defined by how focused the BOT on a particular subject is – is it more generalist or specialist BOT?
|
I see generalist BOTS like Siri, Alexa are more towards high openness to experience vs. specialty BOTS such as task focused – pizza, weather, news bots etc., If Iam creating a Pizza ordering Chabot, I would want it anyhow to ne more task focused vs. openness to experience. |
| Conscientiousness | This trait could be defined by the approach of how the conversations are conducted – is it more open ended vs. menu driven? More the planned, the more the conversation is orchestrated via menus. More menu driven bots in my opinion could be characterized as high in conscientiousness.
Dialogs such as below would relate to more high conscientiousness bots – “Please pick one of the options listed below.” “I don’t understand, Can you pick one of the below” |
E.g. Food ordering chatbots where conversations are very focused and can be conducted via menu driven vs. sales process driven chatbots where its open ended. |
| Extraversion | The way the information is collected – actively or passively. Some bots tend to collect first the information even before proceeding with the business process.
The other aspect is not following a predicted path and more outgoing in the responses and opening up conversations such as ““I see you are from LA…What is the weather like today?”
|
E.g. Sales bot would be more extravert where it can be quite adamant that it needs the information first before any further information could be sent. |
| Agreeableness | Use of words that share emotions and showing a state of agreeableness in responses.
Dialogs such as “I agree…” “Yes, I understand…” “Sorry that you are facing…” |
Customer support BOTS should rank higher on trust, helpfulness and empathize more with the customers without any judgements. |
| Neuroticism | The stable the platform is. The more stable, the less neurotic.
I see there are lots of bots which is not well trained or doesn’t have all flows covered or very rudimentary. All these bots would fall under highly neurotic |
More mature bots which are very consumer focused with voice assisted services like Alexa, Siri that has been trained over many consumers would fit the profile of low neurotic. |
This is my effort towards encouraging users to have a good design and planning around any bot development before deep diving into development. 80% Planning and 20% development is how it should be.