AI will only be sustainable if it is responsible
Published November 4, 2026
- Data & AI
Key takeaways
- Trust will shape the future of AI. More than regulation, distrust resulting from excesses or poorly controlled uses could slow its adoption.
- Responsibility and performance go hand in hand. Clear governance helps limit costs, redundant uses and compliance risks, while supporting economic efficiency.
- AI’s impact extends across all areas. Jobs, skills, data, resources and local communities are all part of the equation.
- A third way must emerge: AI that is useful, controlled and contributes positively. Its long-term viability will depend on its ability to generate more benefits than risks and costs for society.
This op-ed written by Cédric Baecher, Partner Sustainability at Wavestone, was published on September 14, 2026, in Journal du Net (France).
To endure, AI must earn the trust of businesses, employees, investors, local communities and citizens.
Neither responsible nor irresponsible by nature, AI depends on the choices that guide its development and use. AI that is neither useful, controlled nor impactful will ultimately be rejected.
Responsible AI remains an evolving concept. For some, it is an oxymoron given environmental, social and ethical requirements. For others, any additional constraint would hamper innovation in a context of global technological competition. The reality is more nuanced: ethics sets the principles, governance establishes the rules of the game, and responsible AI translates them into concrete, value-creating practices.
From the OECD principles and the European AI Act to initiatives in the United States and Asia, the same ambition is emerging: to build trustworthy AI by combining regulatory frameworks with voluntary commitments.
Responsible AI is not limited to its carbon footprint. It raises questions about its impact on resources, jobs, skills, data and the quality of decision-making. Ultimately, only one question matters: does it generate more benefits than risks and costs for society?
Trust as a competitive asset
Performance and responsibility are not at odds. They are now inseparable. According to the Ipsos’ 2025 AI Monitor survey, conducted across 30 countries, 79% of respondents believe companies should disclose their use of AI. Fewer than one in two trust companies to protect the data they use.
Trust is therefore becoming an economic asset. Ignoring it exposes companies to reputational, compliance, employee relations, attractiveness and technology investment acceptance risks.
Acceptance is becoming critical at the local level
This logic also extends to the infrastructure supporting the growth of AI. While strategic, data centers continuously raise questions about energy, water, land use and local economic benefits. According to the International Energy Agency, global electricity consumption by data centers could more than double by 2030 to around 945 TWh, equivalent to Japan’s current electricity consumption. AI is one of the main drivers of this growth.
The challenge is therefore also social and territorial. Without consultation, value sharing or credible evidence of their impact, these projects risk losing public acceptance. Like any major infrastructure, data centers will need to earn and maintain a “social license to operate.”
