The digital think tank HUB Institute publishes today its HUBREPORT "Future of Artificial Intelligence"The project was carried out in partnership with IBM. This report provides an overview of the state of the art of AI and looks at five key areas where AI is revolutionizing the customer experience, while reflecting on the virtuous potential of the technology beyond the strictly commercial sphere. What are the impacts of AI on the retail, banking or insurance, energy, telecom, industrial and hospitality markets? What are the limits and the meaning to be given to AI if we want a humanized vision and a responsible use of its technologies?
« LArtificial intelligence will be the business advantage of the future. » The PWC report states that the market is in a rush to get AI. "A Revolutionary Partnership: How AI Is Pushing Man And Machine Closer.". Deep learning, machine learning, neural networks, companies and learning processes, the wave of AI is sweeping across many industries and sectors, signalling the advent of a 4th Industrial Revolution.
The AI market is booming and today includes a variety of technologies and tools that facilitate human decision making: virtual agent, facial recognition, image recognition etc... and revolutionize the customer experience and services. The ultimate impact: a true transformation of customer/brand interactions, improved productivity gains and a personalized, seamless and simplified "enhanced" customer experience.
This report, produced in partnership with IBM, aims to provide the keys to understanding and presenting the reality of first adoptions and feedback related to the implementation of AI by brands in the field of customer relationship and processes.
After taking stock of this market, the study details various sectoral applications and then focuses on defining the virtuous potential of AI beyond the strictly commercial sphere.
This HUBREPORT Future of Artificial Intelligence advocates a humanized vision and responsible use of AI, which aims to increase human capacity, not replace it. If technology is neutral, it is indeed people who define its uses.
"We must change our mindset with regard to AI - which is a means - and focus on the goal: the Increase of Human Intelligence, Frédéric-Gérard Leveque, Digital Transformation Director, HUB Institute.
"We are at the dawn of a new era of customer experience that aims to create a new approach to interacting with people. This customer experience will combine natural language, understanding of feelings, visual ability and empathy, says Jean-Philippe Desbiolles, Vice President Cognitive Solutions., IBM Watson France.
PART ONE - GLOBAL STATUS OF IA
In this part, the HUBREPORT focuses on defining AI and how it works, while retracing its history and taking stock of the technologies already in use.
1.1 - What are we talking about?
What is artificial intelligence and where does it come from? A scientific discipline invented in 1955, AI consists of algorithms and neural networks. It is called strong AI when it is applied to reproduce the functioning of a human brain, and weak AI when it focuses on mastering a single task. AI is powered by data, which is processed by a discipline called machine learning.
In summary, AI is not a technology in its own right but a set of technologies and tools, which allows data to be transformed into intelligent actions....
1.2 - The AI is already there
AI is already present in our daily lives through robots, chatbots and connected objects, including the voice speakers that have enabled the emergence of V-commerce, the flagship trend of 2018. The Internet of Things connects everything to AI and is at the source of all current innovations: smart city, autonomous vehicle, telemedicine, home automation...
1.3 - The AI rush
If AI is currently growing exponentially, it is because it is benefiting from an intensified acceleration, which is reflected in unprecedented fundraising and financing: the amounts raised by AI startups have tripled between 2010 and 2016. In addition, AI promises increased corporate profitability, with an overall forecast of +38 % in 2035 and an impact on the profits of companies in all sectors.
An unprecedented increase in funding: Between 1993 and 2017, 1980 fundraising events were held on five continents to support projects integrating artificial intelligence. Between 2010 and 2016, the amounts raised by start-ups in the sector have tripled, from $0.6 billion to $1.8 billion.
Geographically, these fundraising events were held for 62% in the US, 7% in the UK, 3% in both Germany and France, 4% in Israel, 3% in India, 3% in Canada, with the remaining 15TP1T in the rest of the world.
The development of AI would represent an economic growth of $14 trillion in gross value added in 16 industries and 12 countries.
1.4 - A fast-growing but immature market
Since 2012, we are witnessing a real acceleration in the race to buy startups as a vehicle for acquiring skills. Objective: to have the software and hardware arsenal to accompany and integrate the technological evolutions finally made possible thanks to artificial intelligence. Quarter after quarter, more than 200 acquisitions were made in this field. (Source : The State of Artificial Intelligence(CB Insights, 2017).
The market is still dominated by the technological giants, mainly American and Asian. Google made 12 acquisitions of AI startups between 2012 and 2017, Apple 7, Facebook 5, as well as Intel and Microsoft.
IBM has made three acquisitions to broaden the offer of its IA Watson :
- March 2014: acquisition of AlchemyAPI, a text analysis and image recognition API
- May 2014: acquisition of Cognea and its virtual assistants
- April 2015: acquisition of Explorys in the field of health data.
In France, the ecosystem is being built up thanks to excellent schools and recognised research centres, but suffers from a silo culture. While all areas of activity are concerned, investment is very uneven across sectors. A study by Spiderbook covers 500,000 companies internationally and highlights an inequality of investment in AI according to the different sectors of activity.
The industries that invest the most in artificial intelligence are those most directly concerned: digital service companies and software publishers (32%). Conversely, other sectors still invest little (from 0.92% for entertainment to 8.78% for the Internet and 4.19% for telecommunications).
1.5 - AI revolutionizes the customer experience
A PwC study reveals that artificial intelligence will be the business advantage of the future. It transforms and enhances the client experience by boosting client engagement, decision support and creativity. The number of possible applications is staggering.
Which AI-based solutions have the greatest impact on your business?
In the study conducted by PwC in April 2017 among 500 decision-makers, virtual PDAs are considered to be the AI solution most likely to impact the activities of decision-makers. In addition, the study shows that 63% of consumers believe that AI will help solve the complex problems of modern society.
How does AI improve the customer experience?
- By creating a new form of enriched customer interaction that is both simplified and personalized and in natural language.
- By revealing hidden correlations to marketers, AI finally enables brands to deliver a truly individualized experience to their customers. How does it work? Understanding structured and unstructured data and continuous learning makes it possible to hyper-customize content and offers.
- By offering product and service recommendations. The transparency and the level of trust associated with the recommendations help to establish a relationship of trust between a customer advisor and his client.
PART TWO - SECTORAL APPLICATIONS
In this second part, the HUBREPORT focuses on five sectors transformed by AI, through an analysis of the issues at stake, case studies of use by large groups to highlight best practices, and a sourcing and a startup presentation to follow.
2.1 - Banking and Insurance
Issues to remember: in this sector, AI transforms not only the end customer experience but also that of employees with virtual assistants that provide them with personalized recommendations for each customer.
Crédit Mutuel's use case: Crédit Mutuel uses Watson, IBM's AI, to analyse the urgency of e-mails received and to answer questions asked in natural language on the basis of a mass of information. This "super trainee" has increased the response speed of the advisers by 60%.
Other use cases: Orange Bank and the Royal Bank of Scotland use the Watson cognitive platform as a virtual advisor directly to customers, guaranteeing 24/7 availability.
Personetics, the startup to watch: the Israeli company is developing a turnkey solution for banking players with real-time behavioral analysis.
2.2 - Distribution
The issues at stake: the ability of AI to predict demand, automate certain tasks and deliver a hyper-customized customer experience makes it a technology that responds particularly well to the challenges of the distribution sector.
The use case of Bazarchic: the company uses AI to personalize its customer path and target its database when sending its newsletter.
Other use cases: Deliveroo and 1-800-Flowers.com also rely on AI to improve their services, whether to optimize delivery or to advise customers.
Cognitive Matchbox, the start-up to watch: founded in 2016, this French start-up specializes in AI solutions based on connecting customers and agents. Its added value: offering a better understanding of the personalities and emotions of customers.
2.3 - Telecoms and services
Issues to remember: in the telecoms sector, AI enables a customer's management costs to be lowered, service quality to be improved and new services with high added value to be created.
Bouygues Télécom's use case: for the operator, AI makes it possible not to mobilise advisers to respond to the most common problems, which results in time savings for employees.
Other use cases: Telefonica and Softbank have also invested in AI to expand their services.
Sigfox, the unicorn to follow: created in 2011, the French IoT specialist has built a global wireless network connecting 2.5 million objects to the web in 45 countries, relying solely on nearby, low-speed power sources.
2.4 - Industry 4.0
Issues to remember: Industry 4.0 is the realm of cobotics, where robots, machines and humans collaborate to gain agility and product customization.
The SNCF use case: AI enables the SNCF to improve its maintenance and reduce its operating costs while optimising its customer service.
Other use cases: Schneider Electric manages a solar farm in Nigeria remotely using AI, while Bosh has created a self-training system for operators using a tablet and connected wristbands to control and guide their movements.
Intesens, the startup to follow: the company creates connected maintenance solutions for industrial applications.
2.5 - The hotel industry
The issues to remember: Automate certain tasks and rethink processes to make the customer experience more fluid and personalize the service.
The AccorHotels use case study: the hotel group uses "smart pricing" software to establish its rates.
Other use cases: In Japan, the Henn-Na Hotel is fully robotized while many other hotel chains integrate automatons.
Victor & Charles, the startup to watch: created in 2014, the French start-up offers hospitality professionals to know in advance the tastes, habits and "potential" of their customers via its AI solution based on the analysis of public numerical data. This translates into personalized e-mailing campaigns, surprises prepared for certain customers or the creation of tailor-made stays.
PART THREE - IA FOR GOOD
This last part raises the question of the limits and meaning to be given to AI by advocating for a humanized vision and responsible use of AI technologies.
3.1 - Should we be afraid of AI?
While many experts are concerned about technological developments and predict a replacement of humans by AI, it is important to put the intelligent dimension of AI into perspective. AI has neither ethics nor conscience. It is our responsibility to use it wisely. The disruption of employment will result in a transformation of the market and a shift of human skills towards "soft skills".
The short-term impact of artificial intelligence depends on who controls it. And, in the long term, whether it can simply be controlled.
Stephen Hawking, theoretical physicist "
Is artificial intelligence really intelligence?
The "intelligent" dimension of AI needs to be put into perspective. AI has a very mathematical intelligence, based on rigorous algorithms that leave no room for improvisation. This "intelligence" differs from human, biological and dual intelligence (depending on both reason and emotions). The artificial dimension takes on its full meaning here.
Moreover, AI only draws its intelligence from a database, which, even if it is extremely large, serves as a limit to the machine. Its discourse will always be factual, unlike humans.
Voluntary alienation by the machine is therefore highly implausible. AI has no ethics or conscience. It is our responsibility to use it wisely.
Intelligence is not only logic or mathematical reasoning. AI remains a cold world, which experiences neither happiness nor unhappiness, nor envy. AI has no friends or lovers. They cannot be hypnotized or exasperated. They don't feel pride after beating a human at chess, let alone their lack of instinct, empathy or common sense. The fundamental error, which comes mainly from across the Atlantic, is to try to separate the body from the mind.Didier Schmitt, Former member of the Bureau of European Policy Advisers (BEPA)".
For some, AI disrupts employment. We often hear that AI will kill jobs and compete with humans. According to Gartner, one third of the world's jobs will be replaced by algorithms and artificial intelligence technologies by 2025.
AI is complementary to humans. AI increases skills and makes work easier either by increasing performance or by relieving humans of ungrateful and low value-added tasks.
Like any innovation (steam engine, electricity, computerization,...), AI will make a certain number of industries and trades evolve... cutting some jobs but creating others.
AI will transform the labour market by freeing up employees' time to focus on higher value-added tasks. According to a report by Dell and the "Institute for the Future", 85 % of the jobs of 2030 do not exist today!
READ IN UP' : The ten jobs of tomorrow that don't exist today
For others, AI is not ready to take human jobs:
Attention must be paid to the difference between the ability of machines to perform certain tasks and their ability to do certain jobs. If you take the example of the autonomous car, imagine a vehicle that would drive by itself in 99 % of the cases: it will always need a driver. Now imagine a car that is autonomous in 100 % of cases. The difference may seem small, but the impact on the labour market is enormous: either you need a human driver or you no longer need one. And it will take longer than you think to get to the point where machines do all the work and really disrupt the labour market.Nick Bostrom, Philosopher and Professor at Oxford."
AI brings benefits in terms of productivity and growth. For example:
- Retail: Increased online sales of 30% for retailers through the use of dynamic pricing and customization.
- Manufacturing: Acceleration of 30% of production for working-class industries thanks to machine learning.
- Electrical energy: A potential reduction of 10% in electricity costs through the use of deep learning to anticipate demand.
- Health: A potential reduction of 10% in health expenditure with tailor-made treatments thanks to the learning machine.
AI pushes humans to reinvent themselves: It is useless for humans to try to fight AI in certain fields, such as mathematics or pure knowledge of facts. On the other hand, the human still has a role to play in an "integrated AI" working world since the differentiation will be made precisely on the human and more particularly on his creative and emotional capacities.
The potential of technology is infinite but infinity is not enough to make a truth, it is only a field of possibilities. By developing according to our reactions, technology never ceases to question us like a mirror out of ourselves. A revealer that constantly pushes back our convictions about who we are and how we function and reinvent ourselves.Benoît Raphaël, Expert in digital innovation and media "
3.2-Ethics, a key issue
Now that the technologies are working and opening up the horizon further and further, it is important to ask the question of limits.
Limits are no longer technological, they must be human and AI developers must ask themselves the right questions. We don't do AI for AI's sake but to achieve a goal. There is a real issue of overcoming human bias, which will only be possible by installing transparency around AI.
The researchers, engineers and entrepreneurs who contribute to the design, development and marketing of AI systems are expected to play a decisive role in the digital society of tomorrow. It is essential that they act in a reasonable manner, taking into account the socio-economic impacts of their activities. The objective is simple: to force AI developers to ask themselves the right questions at the right time.Cédric Villani, Mathematician and deputy from Essonne, in charge of Artificial Intelligence "
Which AI and why? Having a goal and a strategy. We don't do AI for AI's sake but to achieve a goal.
AI and technology are neutral and have no ethics or conscience. It is up to us to decide whether we want to use them for good or evil.
Following an international conference organized by the Institut du Futur de l'Humanité, on January 12, 2015, the Institute published an open letter to encourage the research community to focus not only on developing the capabilities of artificial intelligence, but also on its benefits to society and the building of safeguards.
Several visions of AI are developing, with a debate on the degree of "awareness" that an AI should have. We talk aboutstrong AI or fromlow AI. It is now important to define a framework for development and to anticipate possible abuses in order to prevent them.
Le report The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigationpublished on Tuesday, February 20, 2018 by 26 AI specialists, provides recommendations to avoid or limit the main hazards.
A responsibility of each one to put limits to what we build: For Kate Crawford, author and researcher on data systems, machine learning and artificial intelligence, civil disobedience could frame the uses of AI.
It refers to the French engineer René Carmille who sabotaged the tabulators used by the Nazis to locate French Jews. AI engineers must therefore consider which lines should not be crossed by their technology. Are there things that should not be built?
AI to overcome human error and bias:
- AI represents an ambition: that of transcending humanity and forming machines that would be "better" than humans. In this case, we are facing a real challenge to correct our human biases. With the learning machine, AI risks reproducing the same biases as those of humans.
- An AI who would analyse a company's recruitment methods using machine learning could reproduce the same gender and race bias.
- An AI built only by white western men risks reproducing "naturally" certain sexist and racist biases. Faced with this issue, many associations have been formed, such as Women in Machine Learning or Black in AI.
A crucial issue: transparency. Artificial intelligence systems have become very opaque and do not explain decision-making processes. AI cannot be a black box from which one gets an answer that cannot be explained. It is essential to know how and by what means AI works in order to be able to control it, for example to understand what element of analysis led an AI to propose a particular medical diagnosis.
At the World Economic Forum in Davos 2018, Virginia Rometty, IBM CEO, called for "a new era of data accountability" characterized by transparency around AI :
With regard to the new capabilities of artificial intelligence, we must be transparent about when and how it is applied, who has implemented it, with what data and in what way. We need to explain why its algorithms make this or that decision. Principles of data stewardship and transparency in AI are essential.
(Source: Cognitive Era: the 3 essential principles according to IBM to guarantee transparency and trust IBM, 2017)
At the end of 2017, IBM joined forces with MIT to create a joint research laboratory, the MIT-IBM WATSON AI LAB whose priority is the control and transparency of artificial intelligence.
This collaboration is based on four axes:
1. advancement of basic algorithms to make development processes more transparent,
2. the development of new software and hardware,
3. AI prosperity - societal impact analysis,
4. Potential applications of AI in society.
3.3 - AI represents hope
AI can be "for good. and work for the common good as a means to achieve the sustainable development goals set by the United Nations. The AI XPRIZE competition sponsored by IBM Watson thus puts in competition teams working on AI applications that help to meet major humanitarian or sustainable development challenges.
Europe and particularly France have the opportunity to establish themselves as players in an AI that would focus on ethics, as President Macron pointed out at the summit. AI for Humanity following the delivery of the Villani Report "Making AI Meaningful".
AI is a sector dominated by the United States and China thanks to their digital giants with huge research and development resources. But Europe is not out of the picture and has a real opportunity to establish itself as one of the leaders in the sector.
The GAFAMs, and now their Chinese counterparts, are already living off the computer exploitation of our personal data. The AI will increase their grip even more. The possible flaws in AI applications are legion and cybersecurity has great days ahead of it. Other vulnerabilities await us, such as the transition from 'fake news' to mass manipulation. However, a new era is inevitably opening up: there will be a 'before' and an 'after'. This is why the continent of the Enlightenment must wake up. It must become an actor that imprints its specificity and ethics, and thus avoid AI becoming a real Pandora's box.Didier Schmitt, Former member of the Bureau of European Policy Advisers (BEPA)".
There is a French ecosystem with actors who are asking themselves the question of the framework to be given to AI and the limits to be set. France can emerge as a third way which would be centred on ethics.
In January 2017, at the end of his term of office, François Hollande launched the FranceAI initiative with the former Secretary of State for Digital and Innovation, Axelle Lemaire, and Thierry Mandon, former Secretary of State for Higher Education and Research. A report with strategic recommendations was submitted to the government on 21 March 2017.
Edouard Philippe's policy statement of 4 July 2017 made artificial intelligence one of the priorities of the Macron government, with three key commitments:
1. Leveraging our talents :
- Establishment of a national programme for AI coordinated by Inria, in the form of institutes throughout the territory.
- Doubling the number of students trained in AI.
- Fluidification of public/private exchanges: a researcher will be able to divide his working time.
- Investment of €1.5 billion for the development of AI.
2. Gathering our strength:
- Open up centralized databases to allow players in each sector to use them to innovate.
- This policy of openness must respect the user's privacy and be accompanied by a European framework for the protection of personal data.
3. Establishing an ethical framework :
- Transparency of algorithms.
- Creation of an international group of experts on AI to organize independent global expertise.
- Create education programs that put transparency and fairness at the heart of training.
Why not enjoy unlimited reading of UP'? Subscribe from €1.90 per week.
International players are investing in AI in France. After Google and Facebook, the Samsung group and IBM announced massive investments in France at the end of March 2018.
Samsung will establish its 3rd largest AI research center in the world in Paris or Saclay.
IBM will recruit 400 AI experts in France within two years. Some of these engineers, developers, researchers and data scientists will be based at the Paris-Saclay science and technology centre.
Technology with the values of France is technology by man and for man. To develop artificial intelligence in accordance with these values, IBM will create 400 new jobs in France.
Anything to add? Say it as a comment.
[…] AI, at the service of the augmented company […]