Societal Responsibility of Artificial Intelligence

Towards an Ethical and Eco-responsible AI

Jerome Beranger, Jérôme Béranger

• Publisher: Wiley-ISTE
• Year: 2021
• Print ISBN:
  9781786306944
• Edition: 1
• eBook ISBN:
  9781119831785
• Topic: Technology & Engineering
• Subtopic: Electrical Engineering & Telecommunications
• Index: Technology & Engineering

The digital world is characterized by its immediacy, its density of information and its omnipresence, in contrast to the concrete world. Significant changes will occur in our society as AI becomes integrated into many aspects of our lives.

This book focuses on this vision of universalization by dealing with the development and framework of AI applicable to all. It develops a moral framework based on a neo-Darwinian approach - the concept of Ethics by Evolution - to accompany AI by observing a certain number of requirements, recommendations and rules at each stage of design, implementation and use. The societal responsibility of artificial intelligence is an essential step towards ethical, eco-responsible and trustworthy AI, aiming to protect and serve people and the common good in a beneficial way.

1
Societal and Moral Questioning Around AI and Its Ecosystem

From autonomous cars to facial or voice recognition, artificial intelligence (AI) has developed and structured itself in a spectacular way over the last 5 years and is now part of our daily life and close environment. The widespread use of algorithmic applications feeds our imaginations, hypnotized by the promise of a better world, where the computing power of machines could reduce or even eliminate illnesses, accidents and crimes. At the same time, a growing doubt about AI is beginning to develop, portraying the technology and its exponential progression as a potential danger to the survival of humanity. Indeed, by the end of 2017, entrepreneur Elon Musk – the charismatic head of Tesla and Space X, among others – believed that efforts to make AI safe had a 5–10% chance of success. In doing so, he reaffirmed his earlier prediction that there was a risk that something very dangerous was going to happen within the next 5–10 years. At the heart of the concerns is technological singularity, a concept that predicts a runaway technological progress that could lead to the advent of a superhuman AI that would have autonomous capabilities to improve and evolve.

Moreover, there is no development of AI without the exploitation of gigantic volumes of data (Big Data). Indeed, we have to keep in mind that the computer without information can neither learn nor automate human action. AI stores the information that humans decide to give it. Failing to limit the collection of data, it would be appropriate to control its use and protection. As a result, the acquisition, storage, consumption and management of this Big Data are two decisive requirements for any contemporary society. So, with the digital revocation of AI and Big Data, all businesses and organizations have become aware of the potential that lies before them.

Now they want to highlight this relevant information and take full advantage of it. But how do you leverage accessible information while ensuring that you have a high-performance ecosystem in place to store, analyze and develop it? The emergence of algorithmic systems also raises the anxiety of a world guided and controlled by digital logic. To what extent can we leave the control of our contemporary societies to algorithms and those who design them? How can we guarantee the confidentiality of our private lives from the growing appetite of machines fed by the collection of personal data? How do we prepare for the upheavals and consequences that AI will bring about in many professional sectors? These are all questions that will be the subject of reflection in Chapter 2.

In order to pragmatically apprehend and understand these issues centered on human dignity and its fulfillment, we introduce four challenges and perspectives offered to society by AI (Floridi et al. 2018):

• – whom can we become (autonomous self-realization): AI can assist in self-realization, i.e. the ability of individuals to develop in terms of their own characteristics, interests, potential abilities or skills, aspirations and projects;
• – what we can do: AI allows us to improve and multiply the possibilities of human representation. Responsibility is paramount, given the kind of AI we develop, how we use it, and how much we share with all of its advantages and benefits. AI applications could help, if designed effectively, amplify and strengthen distributed and shared moral systems;
• – what we can achieve (individual and societal abilities): if we rely on the use of AI-related technologies to increase our work capacities, we will be able to delegate certain tasks and especially decisions concerning autonomous systems that must remain at least in part subject to supervision and human choice. It is, therefore, becoming essential to find a balance between, on the one hand, the pursuit of the ambitious prospects and opportunities offered by AI to improve human life and what we can achieve, and, on the other hand, to ensure that we remain masters of these major developments and their impact on human society;
• – how we can interact with each other and with the world (society, cohesion): AI can go a long way in dealing with such complex coordination, supporting more societal cohesion and collaboration, without undermining human dignity and without eroding human self-determination.

1.1. Use cases of AI

Now, AI applications are invading all sectors of activity and professional spheres of the company. Algorithmic uses are tirelessly multiplying and diversifying a little more every day. In each case, AI can be used to enhance human nature and its performance, creating actual opportunities that must be seized and well used (see Table 1.1).

Table 1.1. AI use cases by industry

Industry	AI use cases
Cities and local authorities	Increase user access to public services: – free up agent time by freeing them from repetitive tasks; – guarantee universal access to public service by breaking down the language barrier.
	Simplify citizens’ lives and experiences: ensuring the efficiency of shared services.
	Optimize the management of the public budget: make the funding coincide with the actual consumption of goods.
Education	Better meet the needs of students: – prevent school and academic dropout; – support students outside of the institution; – support students in their choices.
	Transform learning: – promote learning to read; – propose personalized courses.
	Assist teachers: – allow teachers to spend as much time as possible with their students; – offer teachers feedback on their courses.
	Change the report to information.
	Streamline registration procedures.
Banking and insurance	In customer relationships: – develop commercial relationship; – save time; – speak the same language as the client.
	Reduce risk and fraud: fraud identification and anti-money laundering.
	Create new business models: – combine different businesses through data aggregation; – manage large numbers of investment simulations; – the creation of new products to support healthcare professionals in their daily practices.
Health	AI for better public health: detection and treatment: – revolutionize medical imaging; – harmonize care processes and support the doctor in their application; – generate alerts and reminders to healthcare professionals and/or patients; – review therapy and care planning; – recognize medical images and interpret them (radiology, ophthalmology, dermatology, etc.); – assist in paramedical care (paramedical humanoid robot); – assist in medical decision-making and establish predictive analyses via a diagnostic assistant; – enable communication interfaces between patients and healthcare professionals via a conversational agent (chat-bots) (conversational oncology); – monitor patients in real time and adjust their treatments to their individual situation; – make an earlier and more accurate diagnosis; – access to new knowledge; – improve the flow of the city hospital route; – reduce costs and pool resources.
Retail	Lead the customer to the store: – multiply access points: voice; – multiply access points: images.
	Transform the buying journey and improve customer relations: – make the customer’s journey through the store more fluid; – provide an interactive customer experience; – adapt products to customers’ needs and desires; – use facial recognition for various actions toward the customer; – offer customers the assistance of a personal digital assistant (PDA).
	Empower employees to do more by being more efficient: – facilitate the maintenance of in-store shelves; – alert employees when a customer needs them; – proposals targeted to the user. Optimize operations: – optimize the organization of stores and promote sales; – use customer data; – optimize inventory and inventory management; – optimize the delivery of orders and reduce costs.
Manufacturing industry	Optimize the production chain: – optimize the flow of production lines; – optimize the quality of the production lines.
	Improve the maintenance process: – predictive maintenance; – facilitate the work of maintenance agents.
	Strengthen employee safety.
	Make relations with suppliers more fluid.
	Obtain a better knowledge of the customer in order to better serve him or her.
Corporate Finance – CFO	Make sophisticated and reliable predictions: – obtain forecasts of the company’s financial data and guide strategic decision-making; – determine employee bonuses.
	Free up human time (day/human): – answer questions from business teams; – manage the billing process; – check expense reports.
Marketing	AI for creative campaigns.
	Evaluate the performance of campaigns with a new level of accuracy.
	Promise a personalized experience.
	Constantly improve based on real-time feedback.
	A fluid client interface thanks to cognitive services.
Sales manager and customer relations	AI at the heart of the customer journey: – detect buying signals; – better manage the pace of customer interactions; – in store, detect abnormal behavior or optimize sales staff interventions; – automate information retrieval and call center operations.
	Gain time and efficiency in day-to-day work.
Human resources (HR)	Improve the recruitment process: – attract candidates through language analysis; – facilitate the application process; – identify relevant candidates.
	Giving employees the means to develop: – make interactions more fluid; – offer adapted and personalized training; – enable employees to find their next job with the company.
	Relieve HR managers of certain tasks: – contribute to risk and compliance analysis; – predict recruitment needs.

---

CHATBOTS.–

A software robot that can talk with an individual or consumer through an automated conversation service that can be carried out through decision trees or by an ability to process natural language.

---

Finally, we can envision three logics or ways in which AI will impact employment and the field o...

Table of contents

• Cover
• Table of Contents
• Title Page
• Copyright
• Acknowledgments
• Introduction
• 1 Societal and Moral Questioning Around AI and Its Ecosystem
• 2 The Ethical Approach to AI
• 3 Ethical Framework Associated with AI
• 4 Anticipation Around Artificial Consciousness
• Conclusion
• Appendices
• Appendix 1 Ethical Charter of Using AI in Judicial Systems and Their Environment
• Appendix 2 Practical Recommendations of the CNIL Regarding the Ethics of Algorithms
• Appendix 3 OECD Recommendation on AI
• Appendix 4 Questions Concerning the Application of Ethical Standards
• Appendix 5 CERNA Recommendations on Machine Learning
• Appendix 6 Reasons for a “Digital Divide”
• Appendix 7 Holberton–Turing Oath
• Appendix 8 Report Proposals: “For a Controlled, Useful and Demystified Artificial Intelligence”
• List of Abbreviations
• References
• Index
• End User License Agreement