AI: digital responsibility and the war for talent

During the Annual Meeting of the G9+ Institute on March 18, 2024, Ana SEMEDO, vice-president of Mines digital, opened the event with a captivating introduction. She highlighted the importance of the impact of Artificial Intelligence (AI) on our work, our ethical values, and how businesses and individuals can navigate an ever-changing technological landscape. It raised critical questions about digital responsibility and how companies can attract and retain AI talent.

Ana Semedo also presented the theme of the evening “ Artificial Intelligence, digital responsibility, war for talent: insoluble equation? Or winning trio? », highlighting the objective of exploring whether these challenges represent an impossible equation or if, on the contrary, they can lead to a synergy stimulating innovation and promoting sustainable growth.

Laurence Devillers, Professor of AI at the Sorbonne, then took the floor for the introductory keynote. She set out to demystify AI, particularly predictive and generative AI, and addressed issues of digital accountability. She explained how generative AI systems, like ChatGPT, learn and produce new data, emphasizing that these systems, while impressive, do not possess an understanding of the world comparable to that of humans. She stressed the need for an educational approach to better understand these technologies and warned of potential risks, including manipulation of information and generative hallucinations. Laurence also discussed the challenges posed by English dominance in training data and called for a greater transparency and ethical standards in the development of AI.

Together, these presentations laid the foundations for an evening of deep reflection on the future of AI, its ethical management, and its integration into the social and economic fabric. The discussions made it clear that the aim was not only to discuss current challenges, but also to collectively think about how these technologies could be used responsibly and beneficially for society.

Introduction by Ana Semedo, vice-president of Mines digital

Surprised by AI

Ana Semedo expressed how the acceleration of artificial intelligence (AI) has taken many people and organizations by surprise, leading to something of an unexpected whirlwind. This acceleration has raised many questions about the impact of AI on individuals (“me and my talents”), the environment, and society as a whole. This reflection highlights how quickly AI has become an integral part of our daily lives and the challenges it poses in terms of adaptation and understanding its long-term implications.

Ecological AI

Ana Semedo highlighted the importance of considering AI from an ecological perspective, both for humans and for the planet. She discussed the initiatives of the G9+ Institute in organizing a series of conferences and works focused on ecological AI. For example, the week preceding the meeting, the G9+ Institute presented a study carried out with 45 economic players on the impact of AI on their ecological trajectories, highlighting the process of reflection on sustainability and environmental responsibility in the development and application of AI.

Vision of Artificial Intelligence

In her talk, Ana Semedo shared her vision of AI, where she sees both artificial intelligence and augmented intelligence, and how these two aspects of AI intertwine in our lives. It offers a perspective where AI is not just a tool or a threat, but a potential partner in human development, provided we approach its integration with caution and intention.

Invitation to Explore

Ana Semedo encouraged the audience to take a "counter-attack" approach to AI and explore the unknown, referencing the launch by the École des Mines of the "Bauhaus of Transition." This movement aims to work with the unknown by combining research, training, and action, inspired by the Bauhaus movement that transformed architecture and aesthetics. Ana proposes adopting a similar approach to AI and digital technologies, engaging in a movement of exploration and transformation that could prove particularly relevant to our time. This approach calls for a paradigm shift in how we think about and interact with AI, fostering innovation and creativity.

Keynote address “Context and challenges” by Laurence Devillers – Professor of AI at the Sorbonne

The talk by Laurence Devillers, Professor of AI at the Sorbonne, addressed many fundamental and critical aspects of artificial intelligence.

Introduction to Predictive and Generative AI

Laurence Devillers began by differentiating these two forms of AI, highlighting the importance of understanding them in the current context.

Predictive AI, often used in data analysis, uses machine learning to predict future outcomes based on historical data. This form of AI is widely used in financial, healthcare and retail sectors, where it helps make more informed decisions by anticipating future trends and behaviors.

On the other hand, generative AI, a relatively recent advancement, pushes the boundaries of what AI can create. It not only predicts, but generates new content – ​​whether text, images, music or even code – that has never existed before. Systems like GPT (Generative Pre-trained Transformers) and others Deep learning models are radically transforming creative fields, providing the ability to create innovative works of art, writing and solutions to complex problems.

History of AI

She explored the history of AI from its origins to the present day, from the first conceptualizations of thinking machines in the 1950s to today's explosion of AI applications in almost every sector of society . She spoke of key moments like the Dartmouth conference in 1956, often considered the official birth of AI as an area of ​​academic research, until recent achievements in the development of advanced AI systems capable of deep learning and natural language processing. This retrospective highlighted the significant technical advances made over the decades, while recognizing persistent challenges, particularly in terms of ethics, security and governance.

Development of expert systems

Finally, she discussed the evolution of expert systems, from AI tools designed to emulate the decision-making of a human expert in a specific domain, to modern machine learning and deep learning technologies. Expert systems, popular in the 1970s and 1980s, were among the first real-world applications of AI, offering advice based on a codified knowledge base. However, with the advent of machine learning and, more recently, deep learning, AI has become more flexible and adaptable, learning directly from data without requiring explicit programming by humans. This transition marks a turning point in the development of AI, where systems no longer simply follow predefined rules but learn and evolve based on new sets of data.

The formal neuron

Laurence Devillers highlighted the importance of the formal neuron, a fundamental concept in the field of machine learning and deep learning. The formal neuron is a mathematical abstraction of the biological neuron, simplifying its complexity to allow computer calculations. In the context of AI, a formal neuron takes multiple inputs, each weighted by a certain weight, and produces an output based on the weighted sum of these inputs. This output is then passed through an activation function, which decides whether the neuron fires or not. This simplification of the biological neuron makes it possible to build artificial neural networks capable of learning from data. Laurence emphasizes that, despite this simplification, formal neural models are the basis of spectacular advances in AI, illustrating the contrast between the simplicity of these models and the complexity of the human brain.

Role of Transformers

In her speech, Laurence Devillers addressed Transformers, a model architecture that revolutionized natural language processing (TAL) and generative AI. Transformers use an attention mechanism that allows the model to weight the relative importance of different parts of an input when producing an output. This ability to manage long-distance relationships in text allows them to outperform previous architectures on many NLP tasks, ranging from machine translation to text generation.. The Transformers are here based on models such as GPT (Generative Pre-trained Transformer) and BERT (Bidirectional Encoder Representations from Transformers), who can understand and generate language with remarkable efficiency. Laurence highlighted that the introduction of Transformers marks a turning point in the development of generative AI, paving the way for a wide variety of innovative applications.

Tokenization in Generative AI

Laurence Devillers also explained the importance of tokenization in generative AI processes, particularly in natural language processing. Tokenization is the process of breaking text into smaller pieces, called tokens, which can be words, sentences, or other meaningful units.. These tokens then serve as inputs for NLP models, such as Transformers. Tokenization is crucial because it transforms raw text into a form that algorithms can process, enabling learning from textual data. By segmenting text into tokens, generative AI models can better understand the structure and meaning of language, which is essential for generating coherent and contextually relevant text. Laurence highlighted that tokenization is a fundamental step in AI language processing, making it easier for generative models to manipulate and analyze natural language.

The limits of linguistic models

She then highlighted the challenges inherent in current language models, particularly regarding their ability to represent and understand the real world. Although significant progress has been made in the field of natural language processing (NLP) through architectures like Transformers, these models have intrinsic limitations due to their primarily statistical nature and their reliance on training data. Laurence emphasized that, although these systems can generate and interpret language with remarkable efficiency, their understanding of the world remains superficial. They lack the ability to fully grasp the nuances, cultural contexts, and complexities inherent in human communication. This limitation raises questions about the application and reliability of AI in scenarios requiring a deep and nuanced understanding of language and the world.

Hyper-parameters and model behavior

In her analysis, Laurence Devillers also addressed the importance of hyperparameters in the functioning of generative models, such as the deep neural networks used in generative AI. She explained that parameters such as "temperature" play a crucial role in determining the behavior of these models. For example, the temperature in a text generation model affects its propensity to take risks in its lexical choices, thereby influencing the creativity and variability of the generated text. A high temperature setting may encourage the model to explore less likely options, leading to more diverse but potentially less consistent outputs. Conversely, low temperature promotes caution, with more predictable and often more consistent choices. This discussion highlights the complexity of tuning generative AI models and the importance of these parameters in personalizing the generated results.

Anthropomorphism and AI

Laurence warned us about the dangers of anthropomorphism in perceiving and interpreting the capabilities of AI systems. Anthropomorphism, the attribution of human characteristics to non-human entities, can lead to fundamental misunderstandings about the nature of artificial intelligence. AI systems, despite their ability to perform tasks that appear to require understanding or cognition, do not possess consciousness, intentions, or emotions. Laurence pointed out that anthropomorphism can blur the distinction between algorithmic performance and true cognitive intelligence, which can lead to unrealistic expectations and erroneous assessments of technology. This perspective invites a more nuanced and critical approach to evaluating AI systems, recognizing their abilities without overestimating their level of autonomy or intentionality.

Ethics and Generative AI

Generative AI, capable of creating content indistinguishable from that produced by humans, poses profound ethical questions. Potential manipulation, where generated content can be used to deceive or misinform, requires a clear accountability framework. Who is responsible when AI creates harmful or inaccurate content? Is it the model creator, the end user, or both? The answer to these questions requires in-depth ethical reflection and the establishment of ethical standards and regulations to govern the use of these technologies.

Societal reactions to AI

The rapid advancement of AI has sparked a spectrum of reactions from enthusiasm to distrust. Some are calling for a moratorium on the development of certain forms of AI, such as autonomous weapon systems or advanced generative models, until adequate ethical and governance frameworks are in place. Others, however, see these advances as opportunities for progress and large-scale problem solving. This diversity of reactions highlights the importance of an open and inclusive dialogue about the future of AI in our society.

AI transparency and explainability

Transparency and explainability of AI models are crucial for public trust and understanding. When users and stakeholders understand how decisions are made by AI systems, it builds trust in these technologies. This is particularly important in critical areas like health, justice and finance, where decisions made by AI can have significant consequences on the lives of individuals. Efforts are therefore needed to improve the explainability of AI models, including developing more interpretable AI techniques and establishing transparency standards.

The importance of AI education

Contextualization: in a world where artificial intelligence plays an increasingly central role in various aspects of daily life and the professional sector, understanding its mechanisms, its potential and its risks becomes essential for everyone.

• Public Awareness : Raising awareness and educating the general public about AI is essential to demystify this technology. This includes understanding what AI is, how it works, and in which areas it is applied. An informed population is better equipped to navigate the implications of AI in their personal and professional lives, and more able to participate in informed discussions about AI regulation and ethics.

• Training of Decision Makers : Decision-makers, including politicians, business leaders, and opinion leaders, must also be educated about AI to make informed decisions regarding its development and use. This includes understanding the benefits of AI for society and the economy, as well as the ethical, security, and privacy challenges it raises.

• Educational Curricula : the integration of AI into educational curricula, from primary schools to universities, is a crucial step in preparing future generations to live and work in a world where AI will be omnipresent. This goes beyond learning to code to include an understanding of the societal impact of AI and the skills needed to interact with it critically and creatively.

• Continuing Education : In a rapidly evolving field like AI, continuing education is necessary for professionals to stay up to date with the latest advances, applications, and best practices. Online programs, workshops, and courses can help bridge the gap between AI research and its application in industry and society.

• Interdisciplinary Dialogue : Finally, AI education should foster interdisciplinary dialogue among technologists, ethicists, lawyers, and other experts to holistically explore the implications of AI. This encourages a balanced approach that takes into account both technological innovations and human values.

Futuristic AI and human cognition

Laurence Devillers' vision for the future of AI suggests a development towards models capable of learning and acting in more nuanced and contextual ways, similar to human cognition. Embodied learning implies that AI could benefit from a form of virtual or robotic "body," allowing it to interact with the physical environment and learn from that interaction.. This approach could improve AI's contextual understanding, allowing it to better understand the subtleties of the real world which are often difficult to capture with simple textual or numerical data.

Generative AI: potential and limits

As for the potential and limits of generative AI, Laurence Devillers recognizes the transformative power of this technology, capable of creating new and innovative content, from text writing to image generation and beyond. However, she warns of the limitations inherent in these systems. Although they can imitate creativity, generative models rely on the data on which they were trained and may lack true originality or reproduce existing biases in this data. The distinction between the ability to generate content that feels new and the act of authentic creation remains a central challenge for generative AI.

Cultural and linguistic dependence

In the context of AI, the predominance of the English language in training datasets has been a growing concern, raising questions about inclusion and cultural diversity. This dependence can lead to biases in AI models, limiting their applicability and relevance in non-English speaking contexts (e.g. China). Such a situation highlights the need to expand the diversity of training data to include a wider range of languages ​​and cultural contexts, ensuring that the benefits of AI are accessible and beneficial to a more diverse global audience.

Call for frugal AI

The question of the environmental sustainability of large AI models has also emerged as a crucial challenge. Resource-intensive training processes for deep learning models require a significant amount of energy, raising concerns about their carbon footprint. The call for "frugal" AI aims to develop more efficient and less energy-intensive approaches by optimizing algorithms and exploring techniques that reduce the computing power required without compromising performance. This includes research into smaller, more efficient models and more sustainable training practices.

Societal impact of AI

The impact of AI on society is profound and multidimensional, affecting everything from employment and privacy to security and ethics. While AI offers enormous potential for innovation and improving quality of life, it also presents risks that must be carefully managed. This requires careful ethical reflection, careful regulation and active participation of all stakeholders to ensure that the development and application of AI is done in a way that respects human values ​​and promotes the common good .

AI law and French ambitions

In France, as in many other countries, the regulation of AI has become an important topic of discussion. The mention of the AI Act and national ambitions for AI reflects a desire to place France, and more broadly Europe, at the forefront of responsible innovation in AI. This includes establishing regulatory frameworks that encourage the development of ethical, transparent and secure AI, while supporting research and innovation in the field. Regulatory efforts aim to balance the benefits of AI with the need to protect citizens and preserve democratic values.

Towards ethical and responsible AI

Laurence Devillers insisted on the need to adopt an ethical and responsible approach in the development and use of AI. She pointed out theimportance of ethics in the design of AI systems, to ensure that they are used in ways that respect human rights and promote social well-being. This involves close collaboration between researchers, developers, legislators, and civil society to create a regulatory framework and ethical standards that guide the use of AI while respecting the fundamental principles of justice, equity, and respect. of private life.

Round Table n°1 “AI and the war for talent: Needs, risks and opportunities”

Roundtable #1 of the G9+ Institute Annual Meeting, entitled “AI and the War for Talent: Needs, Risks, and Opportunities,” was a key moment to explore the challenges and opportunities presented by the integration of artificial intelligence (AI) into our societies and industries. Moderated by Luc BRETONES, honorary president of theG9+ Institute , this session brought together influential voices in the field of AI and talent management, including Axel Cypel, engineer, speaker and author; Cristel Guillain, Head of Transformation & Talent at Groupe BPCE ; Laura Letourneau, Expert Advisor to the Prime Minister; Henri Pidault, CIO SNCF group and author; And Véronique TORNER, President of the professional union number.

Convergence between AI and Talents

The interaction between AI and human talent has been highlighted as a crucial element for the professional and societal future. AI, with its potential for transformation and innovation, must be aligned with human skills and aspirations to maximize its positive impact. This convergence is seen not only as an engine of economic growth but also as an opportunity to increase general well-being..

Importance of Human

The speakers unanimously underlined the importance of placing humans at the center of the development and application of AI. AI should be designed and used to enhance human capabilities, not replace them, focusing on applications that improve quality of life and respond to major societal challenges.

Balance between Efficiency and Ethics

A recurring theme of the discussion was the need to balance the operational efficiencies brought by AI with ethical imperatives. It is crucial to develop AI that is not only efficient but also aligned with solid ethical principles, taking into account social implications, privacy and security issues, and environmental impacts.

Collaboration and Collective Intelligence

Véronique Torner highlighted the essential role of collaboration and collective intelligence in the field of AI, emphasizing that success in this area requires a collaborative approach, both at national and European level, with particular emphasis on open source solutions to promote collective intelligence and transparency.

Skills Transformation

Cristel Guillain highlighted the importance of continuous skills transformation to adapt to the pace of AI innovations, pointing to the need for continued education to enable talent to navigate and thrive in a rapidly changing professional landscape. Cristel Guillain, Head of Transformation & Talent at Groupe BPCE, and Henri Pidault, CIO at SNCF, shared concrete examples of internal training implemented in their respective organizations to respond to the challenges posed by the integration of artificial intelligence in the professional world.

Examples of internal training at Groupe BPCE by Cristel Guillain :

• Professional Retraining: Cristel Guillain described an ambitious target set by the group aimed at retraining 600 people by 2024, not only in the technological field but also because many jobs requiring technology skills remain vacant. The group managed to retrain 500 people, more than half of whom were in fields related to technology and data, such as the roles of Data Owner and Data Quality Manager. Guillain highlighted the importance of data for AI and how these roles help protect and structure data, which is crucial for AI effectiveness.
• Challenges of Retraining: She also noted the challenges faced in the retraining process, including the difficulty of attracting individuals completely outside the technological field. Nevertheless, these initiatives have been successful and have helped retain employees, providing them with rewarding career opportunities and addressing a skills shortage in the market.

Examples of internal training at SNCF by Henri Pidault:

• Selection for Training: Henri Pidault shared that for a recent training program, SNCF received more than 800 applications but only selected 400 based on logic and interpersonal skills tests. This selective approach aims to identify the candidates most likely to benefit from the training and subsequently contribute to the company with their new skills.
• Societal Commitment: Pidault also underlined the societal commitment of the SNCF in these training initiatives. In-house training is not only a strategy to fill technical skills gaps but also a way to ensure that employees whose skills become obsolete can remain relevant in the job market, consistent with corporate social responsibility. business.

These examples of internal training illustrate the proactive response of large French organizations to the challenges posed by AI. By investing in the training and retraining of their employees, Groupe BPCE and SNCF show the importance of developing internal strategies to meet the growing demand for technology and data skills, while highlighting their role in promoting a responsible and inclusive professional transition.

Digital Sovereignty and Strategic Autonomy

Laura Létourneau addressed the issues of digital sovereignty and strategic autonomy, highlighting the need for a global strategy that embraces AI and data sharing to maintain strategic autonomy in a globalized space.

Open Source Challenges

Henri Pidault discussed the importance of open source as a lever for strategic autonomy, indicating that fostering open source solutions is essential to playing collectively in France and Europe.

This roundtable illustrated the complexity and richness of the debates surrounding AI, highlighting the need for a multidisciplinary approach that integrates technical, ethical, and human expertise to navigate the future of AI in a responsible manner. and innovative. Speakers shared a vision where AI, properly framed and developed, offers significant opportunities for the improvement of society, while recognizing the significant challenges that must be addressed to ensure that its development and application is beneficial for all. .

Round Table n°2 “Prospect towards new trajectories for Augmented Talents and Intelligence”

The second round table of the Annual Meeting of the G9+ Institute, entitled “Prospective towards new trajectories for Augmented Talents and Intelligences”, addressed in depth the future of work, education, and society at large. era of artificial intelligence (AI). With contributions from several recognized experts, this session explored the challenges and opportunities presented by the integration of AI in various sectors.

AI and the Evolution of Skills

Laurent Champaney, Managing Director of Arts et Métiers ParisTech – École Nationale Supérieure d’Arts et Métiers , insisted on the need to adapt teaching to the realities of AI. He highlighted the importance of revising educational programs to incorporate AI learning, thereby preparing students for a changing professional landscape.

Ethics and Responsibility

Laurence Devillers, Professor of AI at the Sorbonne, addressed the ethical issues linked to the development of AI. She insisted on the need to create technologies that respect human values, warning of the risks of discrimination and violation of privacy that inappropriate use of AI can cause.

AI, Work and Employability

Alain Roumilhac, President of Manpower Southern Europe, discussed the transformational impact of AI on employment. Focusing on the revolution in skills required in the labor market, he spoke of the need for workers to adapt to the new professional realities brought about by AI.

Alain Roumilhac shared valuable insights into the practical use of artificial intelligence (AI) within Manpower, illustrating the impact of these technologies on efficiency and productivity.

• Example 1: Optimizing ad postingRoumilhac described how Manpower employed AI, specifically ChatGPT, to optimize the ad posting process. Traditionally, writing ads represents a repetitive and time-consuming task for agencies, often marked by syntactic errors and potentially unintentional discriminatory criteria. Implementing ChatGPT into this process helped generate higher-quality ads more efficiently, freeing up employee time for other tasks, while improving the quality of ads posted. This initiative is part of Manpower's broader efforts to increase productivity through AI, while alleviating the daunting aspects of certain administrative tasks.
• Example 2: Order prioritizationAnother use case mentioned by Roumilhac concerns the development of a predictive algorithm intended for order management. This system helps teams prioritize orders based on pre-established criteria, significantly improving operational efficiency. Roumilhac pointed out that although the algorithm demonstrated that teams were 20% more productive by following its recommendations rather than their own intuition, adoption of this tool was met with resistance. These challenges illustrate the tension between the objective benefits of AI-based solutions and the subjective reluctance to change and trust in algorithms. This example highlights the importance of supporting change and raising awareness to facilitate the adoption of AI in business processes.

These initiatives reflect Manpower's commitment to leveraging AI not only to improve operational efficiencies but also to rethink and optimize work processes, while highlighting the human and organizational considerations critical to successful integration. of AI.AI in the Public Sector

Franck Denié, Deputy General Director in charge of IS at France Travail (France Work), provided its expertise on the application of AI in the public sector, particularly in terms of improving services to citizens. He highlighted the potential of AI to optimize the efficiency of public services and respond more targeted to user needs.

Franck Denié presented concrete examples of the use of artificial intelligence (AI) within his organization, highlighting France Travail's initiatives to improve personalized support for job seekers and companies thanks to AI.

• 1: Data Capture ChallengesFranck Denié began by highlighting the difficulties encountered in collecting quality data. Despite the presence of digital tools, users rarely fill out their profiles exhaustively, limiting the effectiveness of AI-based supply-demand matching systems. The need for a new dialogue with job seekers and businesses has been identified as a priority to move away from traditional, structured models of information capture.
• 2: Generative AI for Improved DialogueTo meet this challenge, France Travail is experimenting with generative AI in order to facilitate richer and personalized interactions. The objective is to capture data from any interaction with job seekers to then offer more targeted services tailored to their needs. This approach aims to transform business processes by capturing higher quality data through improved dialogues, allowing more personalized support for job seekers and a better response to the needs of companies.
• 3: Integration of AI into Information SystemsA crucial step in the use of generative AI at France Travail concerns the reinjection of the data captured into the organization's information systems. Denié underlined the importance of this process to improve the quality of matching between job offers and requests, as well as to strengthen the effectiveness of the support offered. This technical challenge raises the question of the effective integration of new interaction modalities into existing infrastructure, illustrating the complexity of taking full advantage of generative AI beyond its innovative aspect.

Conclusion

By focusing on improving data collection and interaction with users, France Travail seeks to optimize its role as an intermediary between job seekers and the labor market. These efforts reflect a proactive approach in adopting emerging technologies to meet the demands of an ever-changing business environment.

Challenges and Opportunities for Businesses

Speakers also reflected on how businesses can leverage AI to drive innovation and maintain competitiveness. Strategic adoption of AI, taking into account societal and ethical implications, has been identified as crucial.

Conclusion

The session concluded with consensus on the importance of collaboration between educational institutions, businesses, and policymakers to navigate the AI ​​era. Optimism about AI's ability to enrich society was palpable, as was recognition of the ethical and training challenges this entails.

In summary, roundtable number 2 offered a rich and nuanced exploration of the implications of AI for the future of skills, highlighting the need for a proactive and collaborative approach to maximizing the benefits of AI while mitigating its risks. The discussions reiterated the importance of an ethical and human-centered vision in technological development, ensuring a future where AI responsibly supports social and economic progress.

Closing keynote by Jean-Gabriel Ganascia on the theme of ethics and responsibility in the field of artificial intelligence

The closing of the Annual Meeting of the G9+ Institute was marked by the keynote of Jean-Gabriel Ganascia, computer scientist, philosopher, and professor at La Sorbonne University, on the theme of ethics and responsibility in the field of artificial intelligence (AI). This intervention, rich in reflections, highlighted the ethical challenges and responsibilities incumbent on digital players in an era where AI is omnipresent.

Jean-Gabriel Ganascia began by questioning the notion of digital responsibility, questioning its essence and the diversity of its implications. This question revolves around several axes: the responsibility inherent in digital technologies themselves, that of the creators and developers of these technologies, as well as that of users. In addition, he discussed social responsibility in the digital age, highlighting the complexity of clearly defining these responsibilities in such a vast and changing context.

To illustrate these words, Ganascia shared a striking personal experience during a conference in China, where digital technology has a considerable influence on daily life. This anecdote highlighted how the deep integration of digital, including the absence of cash transactions, reflects a major societal shift, raising questions about dependence on digital technologies and their impacts on individual freedom and privacy.

The evocation of this experience served as a prelude to a broader reflection on the ethical implications of AI and digital technology. Ganascia highlighted the need for an ethical approach in the development and use of AI, calling for increased awareness of the potential consequences of these technologies on society. This ethical approach requires collaboration between developers, researchers, policy makers and users, to ensure that AI is developed and used in a responsible and beneficial way for all.

Beyond ethics, responsibility in the AI ​​era also encompasses transparency, non-discrimination and respect for human values. Ganascia highlighted the importance of these principles for building a just and equitable digital society, where AI acts as a lever for progress and not as a source of division or exclusion.

In conclusion, Jean-Gabriel Ganascia's keynote encouraged collective awareness of the importance of integrating ethical considerations at the heart of technological development, emphasizing that the future of AI must be shaped by a shared, centered vision. on humans and their values.

Congratulations to Didier Carre president of the G9+ Institute, Valentine Ferréol, Ana Semedo, Eric Bresson and the whole teamG9+ Institute for the careful preparation of such a wonderful annual meeting!

Source : https://www.linkedin.com/pulse/ia-responsabilit%2525C3%2525A9-num%2525C3%2525A9rique-et-guerre-des-talents-luc-bretones-dr9je/?trackingId=S0TM16sBfimEnXbgs8MHcw%3D%3D

To go further: this reflection on ethics and responsibility extends our guide to designing responsible generative AI et our article on algorithmic power and human responsibility.