Practical insights on AI governance, responsible adoption, workflow design, and regulation for high-trust organisations.
35 insights
Featured insight
AI Governance & Regulation
Responsible AI needs legitimate goals, bounded authority, decision-specific evidence and a real way to challenge and correct outcomes, not an approval click.
Read articleArticle 50 is now applicable. Learn which AI interactions and outputs need machine-readable marking or human-facing disclosure, by whom, and when.
Europe now has a published competence standard for professional AI ethicists. Here is how to turn the role into credible governance and evidence.
EN 18286:2026 gives high-risk AI providers a European QMS reference for the AI Act. Connect obligations, controls and evidence without new bureaucracy.
Turn AI and agent changes into owned decisions, completed actions and evidence that controls work, with reassessment before materially changed operation.
Connect AI literacy to the tasks people perform, the evidence they must check and the current EU AI Act duty to support its development.
Give reviewers the evidence and authority to approve, challenge or stop AI and agent actions before they affect people or records.
A practical AI-use policy connects approved tasks, data boundaries, review and escalation. Test whether the team can apply it to real work.
Decide whether an AI supplier fits the work, total cost and agent permissions, with demonstrable controls, usable evidence and an exit route.
Moving from an AI pilot to daily work requires evidence about ownership, review, data and complete effort. Define those conditions before expanding access.
Assess Swiss models such as Apertus by the actual service, data flows, legal coverage and tested tasks, and follow the Swiss AI rules now being drafted.
Classify legal AI by its intended purpose and role under Article 6 and Annex III, then identify the applicable obligations and dates.
Identify the GDPR and Swiss FADP records, assessments and breach procedures that apply to the actual personal-data processing in an AI workflow.
What an AI debt-support prototype can suggest, what research establishes, and how to evaluate assistance without assuming improved financial outcomes.
Treat AI financial coaching for sustainable finance as an exploratory workflow, with evidence checks, clear responsibilities and a scoped legal assessment.
Distinguish stylistic resemblance, copied expression, misleading attribution and training use before choosing a legal response to AI imitation.
Separate human authorship, third-party rights and contracts when reviewing AI-assisted art, with distinct Swiss, EU and US sources.
A bioethics comic is a format to test for understanding and participation. Separate case-study evidence from legal duties and claims of public consensus.
Assess LLM-assisted health explanations alongside current access rights, EHDS phases and evidence about readability and clinical errors.
Explore AI-assisted contract design through proposals and research on comprehension, while checking source obligations and operational meaning.
Robots that infer emotion raise questions about measurement, lawful processing and usable privacy choices. Consent is not the only possible legal basis.
Evaluate legal AI through accuracy, access, complete effort and procedural rights, with evidence about the actual service and applicable duties.
Compare generative contract tools with research on defined tasks, and test jurisdiction, source accuracy and full review effort before relying on a draft.
AI can help draft contract implementation guides. Verify obligations, preserve source links and test the guide with the people who must use it.
Research supports contracts that are easier to understand. Use plain language and AI drafts, and verify meaning, legal effect and comprehension.
Technology-assisted review can help prioritise document-heavy legal work. Its value depends on independent validation, confidentiality and the actual procedure.
AI can support M&A document review. Evaluate missed issues, source verification and total effort before promising faster or broader due diligence.
How legal and professional-service teams can handle overlapping GDPR, Swiss FADP and EU AI Act duties when they use AI in sensitive or client work.
A practical AI governance framework for teams handling confidential work, regulated decisions, and professional accountability.
Evaluate AI tools for legal and professional work through task-specific evidence, data controls, human review and applicable duties.
Evaluate legal research AI through source coverage, multilingual retrieval, verified answers and complete review effort for your Swiss practice.
Evaluate AI-assisted contract work through bounded tasks, original-source checks and the complete effort of review and correction.
How to evaluate pricing and complete review effort when introducing AI assistance into legal work.
Turn Susskind’s future-of-law questions into a testable service strategy, with evidence about client need, quality, full cost and staff capability.
Revisit Susskind’s forecast alongside research on legal tasks, separating measured changes from predictions about jobs, firms and demand.
35 insights
Every article published in this language, newest first.
Start here
Start with a focused conversation about a live AI use case, workflow bottleneck, training need, or governance gap.