Personality Aware AI & Virtual Assistants
Pad ~ Success Coach
Optimising Team Performance and Profitability - with immediately actionable remote diagnostics
Personality Aware AI & Virtual Assistants
Artificially intelligent software programs designed to interact with humans in a conversational manner, have become increasingly popular in recent years. AI-driven virtual assistants such as 谷歌 Assistant, 苹果 's Siri and 亚马逊 's Alexa, to highlight some of the biggest players, now interact with millions of individuals on a daily basis to provide information and lifestyle support that many couldn’t now imagine living without.
These bots are making use of growing pools of personal data to make their conversations more useful, helpful, effective and efficient, and machine learning tools are gathering and analysing data from these interactions to improve the quality of future engagements.
Yet there is a vital piece in this equation that has been left unaddressed and undeveloped.
Remind Me, Where Did You Leave Your Personality?
In any effective relationship between a teacher and student, manager and employee, parent and child, sales-assistant and customer, there is a vital need to account for the individual personalities of the parties involved.?
It’s simply not enough to spout generic, ‘vanilla’ responses if there is to be any kind of meaningful, emotional engagement between people. We’ve all known teachers, parents, friends and mentors who knew exactly how to ‘get on our wavelength’ - and others that really didn’t.
The current state of play involves the use of all kinds of datapoints including user demographics, past search queries, interaction and purchasing history, and many more. All these do help define some important user preferences, but there are very, very few examples of bots demonstrating any kind of ‘personality awareness’.?
[I’m aware that IBM Video Streaming has implemented DISC and OCEAN profiling models to some degree, but I’m keen to hear of others that are doing this effectively if you have knowledge of such developments.]
While huge strides have been taken to humanise the engagement experience, users are already tiring of their virtual assistants acting in their typical, robotic and impersonal manner, which can be frustrating when trying to communicate with them.
There’s a huge opportunity here.?
The Next Frontier
I propose then, that we’re about to see a surge of developments in the area of mapping the core personality traits of the individual human users and providing the virtual assistants with language sets that engage each user the most effectively.
New sets of ‘emotionally resonant language’ will need to be defined for each personality type for #naturallanguageprocessing purposes; words and phrases that align with the individual’s values, needs and preferences and reflect their way of looking at life.
领英推荐
Building the Framework
All this assumes that we actually have a #personalityassessment model that provides an accurate, usable framework with a clear set of definitions, along with a way to rapidly and efficiently capture data on an individual’s key traits.
Once we have a simple tag that defines the user’s personality type, any personality-aware virtual assistant (PAVA) or personality-aware chatbot (PAC) can interact using far more engaging language patterns - making them more enjoyable to converse with and delivering what is ultimately a more successful and satisfactory experience, whether the use-case involves customer services or conversational companions.
[ You may like to review the fabulous work of Hongyu Jiang, Ao Guo and Jianhua Ma in their paper from 2020 titled: 'Personality-aware Chatbot: An Emerging Area in Conversational Agents' - https://www.researchgate.net/publication/346468952 ]
Do You Like My Tone?
We should add that, beyond the use of specific words and phrases, tonality of voice is also very important in conveying emotional content in any conversation. This kind of expression will always be missing from #chatbots , of course, but virtual assistants will soon also learn to deliver their responses in the most effective voice tone for the individuals with whom they are interacting.?
Such dynamics will significantly improve the quality of interaction in cases such as Virtual Fitness Coaches, for example, where voice tone is a key ingredient in effective interaction and the ultimate response and performance of the human user.
Meeting Market Demand
The market demand for more personalised and human-like interactions with our machines is surging forward significantly. #Chatgpt has proven our voracious appetite for human-machine conversation and the pools of knowledge accessible to these kinds of services are leading to extraordinary levels of productivity.?
We choose to read books by certain authors because we like *their* style of writing, their flavour of storytelling, their manner of presentation. But when it comes to education, training and conversation, we want the other party to account for who *we* are as individuals.?
Unfortunately, when it comes to digital communications, this crucial human factor is still missing. It’s not only that we don’t have machines that can effectively vary their content based on our personality types, we don’t even have an effective model that defines personality in an accurate and manageable way - and, if we do, we don’t have a way to capture this data from individuals in an efficient manner. Profiling methods generally require 12-20 minutes of a manual input from each user, or for a machine to crawl and assess a vast amount of accessible social outputs to get even a hint of their leverageable personality traits.
A Universal Dialect
If this is your field of expertise, you may want to reach out to discuss TEAM ME? DIALECT, an approach that meets all these issues head on and provides a simple model of personality that can be used to define clear sets of emotionally-resonant language. DIALECT also provides a method of rapid profiling (90 seconds of user input) for each individual, to ascertain their unique personality profile.
In later articles, I’ll deep deeper into specific use-cases that may be significantly enhanced through personality awareness, including:
Principal Data Scientist @ Adarga | NLP & ML Science & Engineering expert
1 年Thanks Pad for such a thought provoking article. I have a major concern here though, which is potently captured by the movie Her (2013). Do you agree we need to assess the negative impact this can have on emotionally vulnerable people?
Behavioural Neuroscientist ?? OD Consultant | Human centric research and employee needs analysis to inform leadership development, culture change and wellbeing strategy
1 年This is great, Pad. I love the work you do using personality profiles to improve communication, and applying this to AI would definitely take our digital conversations to a different level. No matter how good AI is, it still lacks the human touch and personality - much more to be done in this space to personalise our virtual conversations.