Press release from swissAI dated January 7, 2026
The Problem: AI Implementation Without Oversight or Training
As freelance journalist and MAZ research director Reto Vogt revealed in his investigation for the *NZZ am Sonntag* on January 4, 2026, the activation of Microsoft’s Copilot in Parliament was effectively carried out autonomously by the American technology company—without an official federal resolution, without clear guidelines, without defined responsibilities, and without transparency. The introduction happened, so to speak, through the back door. This created a legal and security policy gray area.
Members of Parliament and administrative staff work with sensitive—and in some cases, confidential—information. Nevertheless, no comprehensive training programs or binding guidelines have been introduced. Those who use AI without understanding how it works, its limitations, and its risks can neither take responsibility nor build trust—neither within institutions nor with the public.
The use of AI is first and foremost a leadership issue
This situation highlights a fundamental governance issue regarding the use of AI. Political institutions must clarify whether, for what purposes, and under what conditions AI will be used. What is the prevailing understanding of digital sovereignty? What data may be processed? Who decides on its use—and who bears responsibility? These questions must be answered at the political level before technical systems are put into productive use. This awareness is lacking in many places.
swissAI recommends the competent and controlled use of AI
“Our direct democracy thrives on trust in government institutions and independence from external influences. AI is a powerful tool that can improve our processes—but only if we ensure transparency, competence, and sovereign decision-making processes,” says Chris Beyeler, president of swissAI.
In this case, swissAI recommends:
• Policy Clarification: Parliament should engage in a broad dialogue to define which applications are permissible and how Switzerland’s digital sovereignty can be safeguarded.
• Capacity Building: An assessment of training needs and needs-based continuing education for Parliament and the administration to understand the risks and capitalize on opportunities.
• Incorporation of expertise: Independent specialized agencies should support the technical and ethical assessment and ensure that the rollout is in line with data protection and Swiss values.
Sovereignty in Dealing with AI
Switzerland possesses excellent expertise in the field of artificial intelligence. This potential must be harnessed—in a coordinated, responsible, and sovereign manner. AI can be a competitive advantage. However, this requires that its implementation be carried out professionally and not left to chance or licensing mechanisms.