AI and Voice Recognition: Trends to Follow Closely

Originally published in Journal du Net [FR] IA et reconnaissance vocale : des technologies à suivre de près

By Baptiste Bouffaut, CTO of Ponicode

“AI is technology’s most important priority and healthcare is its most urgent application”.

These were the words of Satya Nadella, CEO of Microsoft, commenting on the acquisition of Nuance, a pioneer in artificial intelligence (AI) and speech recognition, in April 2021.

Founded in 1992, Nuance produces software that allows you to dictate messages to a computer or transcribe spoken exchanges. The company has significantly influenced research into NLP (Natural Language Processing), the branch of artificial intelligence that enables machines to understand our language and communicate with us.  

In acquiring Nuance, Microsoft is highlighting the real revolution that this technology is undergoing. NLP is all around us; it can be found in a number of everyday tools like Google Translate, Alexa, Siri, and even customer support chatbots. This technology based on machine learning gives software the capacity to analyse the human language, to understand the meaning and then reproduce it. Digital assistants are certainly the best-known application of NLP, but the use cases go far beyond that. The recent democratisation of AI algorithms and NLP models has really boosted research into semantics and natural language.

The use cases of NLP are increasing across all fields

The real innovation of NLP is that it allows artificial intelligence to understand text by taking into account the context. Thus, algorithms using NLP can distinguish the meaning of a word according to the situation in which it is used.

In addition to the mass-market use cases of voice assistants and web chatbots, NLP is already widely used in the analysis of social networks, where it is possible to establish the tone of a post or a blog post with regard to a company, a brand or an influencer.

So-called "Sentiment Analysis" is an approach that is now perfectly mature and effective. In just a few tweets you can find out whether a consumer prefers Coke to Pepsi, even if it is not directly mentioned in the tweets...

In the banking and insurance sector, NLP has many uses: anti-fraud, analysis of documentation to grant a loan, analysis of internal documents for compliance purposes, and of course analysis of markets to optimise investments, NLP, which is still in development on these subjects, is set to become a "commonplace" technology for the sector. In terms of text generation, NLP is being used to generate analysis reports for investors and these algorithms are being tested in some newspapers and on some specialised news sites for writing press articles to comment on companies' financial results and sports matches. 

Numerous NLP business applications are emerging, particularly in the industrial sector where employees have smart assistants to help them in their daily work. For example, Vallourec's technicians use voice assistants to take dictation of measurements made with instruments that require both hands. Similarly, in logistics, smart assistants can offer unparalleled flexibility of use thanks to natural language and can replace voice-controlled systems that turn logistics operators into mere execution robots.

Code is the most spoken language in the world

In computer development, a new approach called "AI on Code" is exploiting artificial intelligence models to analyse computer code written by developers. A NLP model can analyse computer code in exactly the same way as human language. The intention of the developer is extracted by the model. This provides information that can be used to generate relevant unit tests that will assist the developer in creating software. This is mind-boggling when you consider that code is the most widely spoken language in the world. 

And how will you use NLP?

By freeing people from the role of understanding and creating language, NLP makes it possible to imagine a huge number of optimisations in all industries and for all tasks involving language. Access to NLP is still costly, but the barriers to its democratisation are gradually disappearing.

All companies will soon be able to access these technologies and shake up the way procurement, distribution, production and customer relations have been approached until now.

We are already seeing the competitive advantages generated by NLP today. That is why it is crucial to encourage companies to explore the possibilities in this area. It’s possible for all companies to initiate conversation around how NLP can transform their project:

  • how can you improve processes within your company?
  • how can you free your teams from repetitive tasks so that they can concentrate on tasks of high added value?
  • how can you drive your project towards this new industrial revolution? 

These are the questions that NLP can allow us to answer. There is no doubt that this opportunity for our economy will propel our products and services into a new era as the applications of NLP are only limited by our inventiveness.

🇫🇷 For French readers, here is the article published in the Journal du Net

Green blobred blob