Registration is wrong! Please check form fields.

Artificial Super Intelligence – the Future of Speech Technologies and a Sense of Direction for Insurance

Artificial Super Intelligence – the Future of Speech Technologies and a Sense of Direction for Insurance

Artificial Super Intelligence – the Future of Speech Technologies and a Sense of Direction for InsuranceLonger term projections about the pace of artificial intelligence (AI) development into what some scientists refer to as artificial super intelligence (ASI) belong to the realm of futurology at present. Nobody is quite sure what the future holds given the multitude of almost entirely unpredictable factors shaping human history, including the technological part of it. However, one thing is certain –  we can presently see how the market forces are using the momentum created by the three push-factors influencing an extraordinary transformation of the way we work, exchange information, and even do things at our leisure:

  • break-through in developing superfast processors,
  • cognitive systems that can process unstructured data, constituting 80% of all data out there, as well as the knowledge base produced by the humankind in the form of books and other written heritage, and
  • the rise of natural language processing (NLP) technologies based on machine learning.

The urgency is backed up by investment forecasts of stunning proportions. A Narrative Science survey found last year that 38% of enterprises are already using AI, climbing to 62% in 2018. Forrester Research predicts +300% increase in AI investment this year compared with 2016. IDC estimates the AI market will grow from $8 billion in 2016 to more than $47 billion in 2020. Tractica estimates the speech biometrics market exceeding 5 billion USD by 2022. MarketsandMarkets forecast the speech technology sector will be worth 10-12 billion USD in five years.

There is a clear trend and emerging consensus among experts that spoken languages and verbal commands will become the key interface for communicating with machines. But there is an important obstacle for AI developers. The spoken word alone does not convey the entire load of meaning that people perceive in communication with each other. Even if the super-segmental level of the language, e.g. intonation, as well as the emotions recognition, can already be translated into machine readable content with a varying degree of success, there is another domain, which evades technological understanding – the intuitive and largely non-verbal perception of meaning. It does rely on certain elements of a language like the choice of words as indicators of meaning, spontaneity and sequencing of utterances, and supportive body language.

It can be illustrated by an example familiar to virtually all people: we prefer a personal meeting face to face to discuss something important or highly personal. Distance in dialogue as well as any technical medium affects the quality of communication, even if it is the case of a video-chat and larger-than-life crystal clear imagery. We can only understand the other fully, when brother embraces brother – there’s no better way of putting it (with all due respect to sisterhood, which is another issue – gender-sensitive semantics).

It can be argued, of course, that in most communication situations this kind of semantics is less important that purely linguistic carriers of meaning. Furthermore, some emotional characteristics such as a timbre of the voice, its tone and tempo as well as other features that make up the nature of its attractiveness or assertiveness for example can be emulated by computer systems based on copying real human voices and their modulations.

Talking face to face to the other person is meeting with another human being in his or her fullness of being, experience, and uniqueness of personality. This is a hugely important although rather intangible part of human inter-connectedness and mutual participation that cannot be reproduced by manipulating elements of what we know about ways of human communication and isolating areas of the unknown by diverting attention in dialogue with the machine. The big void will naturally continue to be there for many, but will we be willing and able to understand it amid the business rush and over the phone?

Now, imagine a machine finally passing the Turing test with the score that tells us it’s basically human. Consider then that this machine can tap into unimaginably rich knowledge-base of millions of books of literature, psychology, communication etc. with all the versions of the answers ideally suitable for the questions posed. Take it a little further, and imagine a complete awareness of the communication situation, timing, emotional state of the speaker and even his or her medical condition, as well as many other kinds of information that may be extracted from speech and the device being used. We are getting closer here to what is crucial in distinguishing the difference of ASI and its impact.

For most business interactions, especially in the customer service, human-machine dialogue will replace costlier human-to-human interactions without any significant loss of service quality within the next 5-10 years. It will make human-machine dialogue ubiquitous in day-to-day life for purely economic reasons. Certain impact on the labor market and even extinction of some professions is unavoidable, but it will eventually lead to higher quality jobs creation especially in the service industry.

Insurance is one of the industries where the pressure of the incumbent disruption (e.g. new customer service and risk calculation approaches by InsurTechs) will be especially pronounced. In fact, InsurTechs are already making their bets on leveraging the existing data to generate risk insights, develop new capabilities in customer service that make it easy and more fun to deal with them, and finally – offer trusted dialogue-based customer relationships with effectively personalized products and hassle-free delivery.

According to Deloitte, “psychological biases inhibit sales of life and annuities products. More effective messaging tools and products to overcome behavioral barriers should be developed.” This and other similar tasks can be addressed through effective communication with customers and these business processes can be radically strengthened using NLP-enabled artificial intelligence, sentiment analysis, personalization of services achievable through effective speech analytics and other methods.

Thinking about the future today is critical and having an integrated 360-degree view of the full spectrum of speech technologies and their future as a part of artificial super intelligence gives an important sense of strategic direction. Eventually, AI-powered analytics in conjunction with speech technologies such as semantic interpretation and emotion detection will be effectively helping insurance contact centre specialists to be even better at persuasion, social understanding, and empathy – the creative qualities that will most likely remain up to humans to excel in.

It is the right time to proceed from experimenting with NLP-enabled services to a meaningful transformation of business processes in line with newly opened technological possibilities such as those provided by Spitch AG. Thinking strategically and reducing complexity is the right way forward: take small steps first, consolidate; reduce new complexity, take more small steps forward, consolidate; assess performance improvements and repeat. Some leading insurance companies are going to discuss how to do it exactly with their partners: Spitch, Adcubum, Swisscom, ti&m and others focusing on concrete cases ready for implementation. If you would like to attend, please fill out the registration form. Available spaces are limited and filling up quickly, so please, register today.


website in progress