Between fascination and concern, AI sparks both excitement and debate in the workplace.
Capable of transforming practices by efficiently automating certain tasks or supporting activities traditionally reserved for humans, AI also raises major issues: technological maturity, environmental challenges, and social responsibility. As such, the sustainable integration of AI into organizations requires above all an ethical, adaptive, and rational approach.
Adoption still in the exploratory phase
For organizations, the potential of AI is highly promising. However, contrary to the 2022 predictions and the arrival of ChatGPT, the actual adoption of AI in everyday practices remains marginal. Currently, companies are multiplying tests, proof-of-concept experiments, and pilot projects. In project management, for instance, only 5% of employees use AI daily, 36% occasionally, and 48% rarely or never.
How can such a gap be explained? First, by a lack of skills and understanding of AI applications. For successful integration, decision-makers must grasp the underlying issues to develop a clear strategic vision that will lead to sustainable transformation. Employees also need support to overcome resistance and change-related concerns.
The technical challenges are significant. In a constantly evolving environment, what kind of investments should be prioritized? A custom-built solution or an off-the-shelf tool? And then there’s the issue of hallucinations—false or misleading outputs—which, while becoming less frequent, still require specific model training (fine-tuning, prompt engineering, etc.) to achieve adequate reliability.
Environmental and ethical challenges: finding the right balance
The potential of AI and the excitement surrounding it should not obscure its environmental and social impacts. Ecologically, the resources required for AI operations are often criticized. A single ChatGPT query consumes 21 times more energy than a Google search—a staggering difference. On the infrastructure side, the data centers that run AI models require massive amounts of electricity and water for server cooling. Green AI (resource-efficient AI) and AI for Green (AI used for sustainable development) are often seen as opposing approaches, but it is the combination of the two that holds the key to tackling environmental challenges.
Ethically, data governance and social disruption raise critical questions. European regulations (GDPR, AI Act, Data Act, Data Governance Act) frame AI use: protecting personal data and employee rights, preserving individual freedoms, and respecting intellectual property, digital sovereignty, and the working conditions of so-called “click workers.”
These impacts on the planet and society prompt further reflection: What is the appropriate scope of AI usage? How can compliance with legal standards be ensured? How can the quality of training data be maintained?
Regulating AI activities is fundamental. The sovereignty and sustainability of our businesses are at stake.
Expertise: a key factor for mature, sustainable AI solutions
Without expertise, there can be no trustworthy AI. Developing the skills of all stakeholders requires the creation of an AI Center of Excellence (AI CoE). Why? To address three key challenges: team training and awareness, sharing best practices, and incubating new solutions.
- Training to better master AI
Teams must acquire diverse skill sets to address both technical and functional challenges across internal and external projects. Since many functions are involved, a cross-disciplinary and transversal approach is essential to better meet customer expectations.
- Sharing and structuring resources
Beyond talent, resource management is crucial: computers with graphic processors, local and cloud servers. These sometimes-costly resources need to be pooled and prioritized according to the company’s strategic goals.
- Innovating to stay ahead
Experimentation is vital: should we buy an existing tool or develop an internal solution? In the former case, data governance and cybersecurity are critical. In the latter, the OKR (Objectives and Key Results) method allows for the definition of clear goals and performance tracking to continuously adjust the strategy.
Beware of alarmist rhetoric or exaggerated promises—AI should be approached with realism. Starting today, we must structure its use, invest in skills, and work collaboratively. We have the opportunity to turn AI into a driver of competitiveness and positive impact. Let’s make the most of it—wisely.
