There is currently no generally applicable statutory definition of “artificial intelligence” or “AI system” under Swiss domestic law.

Federal materials on artificial intelligence regulation rely on or are closely aligned with the definition of an “AI system” set out by the Council of Europe’s AI Convention. According to this definition, an AI system refers to “a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations or decisions that may influence physical or virtual environments. Different artificial intelligence systems vary in their levels of autonomy and adaptiveness after deployment.”

On its website dedicated to the Competence Network on Artificial Intelligence (CNAI), the Federal Chancellery uses a very similar definition and describes an AI system as “a machine-based system that uses the input it receives to infer explicit or implicit goals and generate output such as predictions, content, recommendations or decisions that can influence physical or virtual environments. AI systems can be equipped with varying degrees of autonomy.”

Switzerland has not adopted a single, standalone national AI strategy. Its approach to AI is embedded in broader digital policy instruments, in particular the Digital Switzerland Strategy and the Digital Federal Administration Strategy. In addition, the Federal Administration has adopted a dedicated strategy for the use of AI systems within the Federal Administration.

The objective defined by the Swiss Federal Council is to regulate AI in such a way that its potential can be used to strengthen Switzerland as a location for business and innovation while keeping the risks to society as low as possible.

In February 2025, the Federal Council decided to focus on the following parameters:

1. The AI Convention of the Council of Europe, signed by Switzerland in March 2025, shall be incorporated into Swiss law, applying primarily to state actors.

2. Where legislative changes are needed, they should be as sector-specific as possible; only key areas relevant to fundamental rights, such as transparency, data protection, non-discrimination and supervision, shall be subject to general, cross-sectoral regulation.

3. In addition to legislation, non-legally binding measures, such as voluntary commitment declarations or industry solutions, shall be developed to help implement the Convention.

To implement these parameters, the Federal Council mandated the Federal Department of Justice and Police (FDJP), in cooperation with the Federal Department of the Environment, Transport, Energy and Communications (DETEC), the Federal Department of Foreign Affairs (FDFA) and other relevant federal offices, to draft a bill with the necessary legal measures to implement the Council of Europe’s AI Convention. DETEC has further been instructed to propose complementary non-legally binding measures. Both the draft bill and the plan for non-legally binding measures are expected by the end of 2026.

The strategy for the use of AI systems in the Federal Administration has been valid since March 2025. It sets out three fields of action: developing skills, earning trust and increasing efficiency. It is implemented through a separate implementation plan, which is updated regularly and carried out using agile methods.

In October 2025, the Federal Office of Justice organised a participatory workshop with around 60 stakeholder representatives from politics, public administration, business, academia and civil society to discuss the incorporation of the Council of Europe’s AI Convention into Swiss law. The outcome of the discussions was summarised in a report and is intended to be considered in the further legislative process.

Switzerland does not currently have any overarching AI-specific legislation. The Federal Council signed the Council of Europe’s AI Convention in March 2025, but the Convention has not yet been incorporated into Swiss law. A consultation draft for the necessary legal measures is expected by the end of 2026 (cf. question 2).

However, Switzerland has adopted several sector-specific guidelines and supervisory expectations:

– AI Guidelines for the Confederation: The Guidelines on Artificial Intelligence for the Confederation were adopted by the Federal Council in 2020 and are reviewed and assessed by the Federal Office of Communications (OFCOM) every two years. They apply to the Federal Administration and agencies entrusted with governmental tasks and aim to provide a general frame of reference and ensure a coherent policy on AI, based on seven principles: (1) putting people first; (2) regulatory conditions for the development and application of AI; (3) transparency, traceability and explainability; (4) accountability; (5) safety; (6) actively shaping AI governance; and (7) involving all relevant national and international stakeholders.

– Strategy for the use of AI systems in the Federal Administration: Switzerland has adopted a dedicated strategy for the use of AI systems within the Federal Administration. It applies to all AI systems used in the Federal Administration, regardless of the specific AI technology used or the purpose it serves. The strategy is based on principles including people-centred and trustworthy use, clearly defined responsibilities, sustainability, reliability and robustness.

– FINMA Guidance 08/2024: The FINMA Guidance “Governance and risk management when using artificial intelligence” outlines supervisory expectations for the use of AI in supervised Swiss financial institutions. It focuses on governance, inventory and risk classification, data quality, testing and ongoing monitoring, documentation, explainability, and independent review.

– AI Code of the Swiss media industry: In 2026, Swiss media industry associations adopted a Code of Conduct for the responsible use of AI systems by media companies. The Code addresses, among other things, user knowledge, protection of democratic processes, personal data and confidential information, transparency, labelling obligations and complaint mechanisms.

Furthermore, there are several existing laws that apply to the development and use of AI systems in Switzerland:

– Federal Act on Data Protection (FADP): The FADP applies whenever personal data is processed. It contains, among others, provisions on transparency, data minimisation, data protection impact assessments, automated individual decision-making and profiling. The Federal Data Protection and Information Commissioner (FDPIC) has emphasised that the FADP applies directly to AI-supported data processing whenever personal data is involved. One interpretive challenge is that AI can increase the risk that pseudonymised or anonymised data may be reassigned to a specific person.

– Federal Copyright Act (CopA): The CopA is currently under discussion in relation to the use of copyright-protected works and performances by AI systems, for AI training and output generation. The issue is not expressly regulated under the current CopA. A parliamentary motion requested that copyright-protected information may be used by AI systems only with the copyright owner’s consent. Parliament adopted the motion in amended form, mandating the Federal Council to prepare the necessary legal basis to better protect journalistic content and other copyright-protected works and performance against AI-related misuse, while preserving Switzerland’s competitiveness in AI research, development and commercialisation. The precise legislative approach has not yet been determined.

– Federal Product Liability Act (PLA): The PLA is formulated in a technology-neutral way and can, in principle, apply to liability issues related to AI-enabled products. However, its application to AI raises interpretive challenges regarding defectiveness, causation and evidentiary issues where damage is caused by complex, adaptive or opaque systems. It is expected that Swiss product liability law will be reviewed considering technology developments and the revised EU Product Liability Directive.

– Federal Product Safety Act (PSA): Swiss product safety law may apply to AI-enabled products where they qualify as products and no specific sectoral product safety rules apply. The Federal Council opened a consultation on a partial revision of the Product Safety Act on 5 June 2026, primarily to strengthen product safety and introduce new rules for online commerce. The consultation runs until 28 September 2026.

– Sector-specific laws: Several sector-specific laws may apply to AI systems. For example, the Federal Act on Medicinal Products and Medical Devices (Therapeutic Products Act, TPA) and the Medical Devices Ordinance (MedDO) may apply to AI-based medical devices or software used for medical purposes. The Road Traffic Act (RTA) and the Ordinance on Automated Driving (OAD) regulate the use of vehicles with automated driving systems. These regimes do not constitute general AI legislation, but they regulate specific AI-enabled applications.

– Other laws: Depending on the use case, other legal regimes may also apply, including contract law, tort law, employment law, criminal law, anti-discrimination law, public procurement law, administrative procedure law, and financial market law. The main interpretive challenges are the application of technology-neutral legal concepts to AI systems, including transparency and explainability, allocation of responsibility, human oversight, evidentiary issues, discrimination and bias, cybersecurity, and the treatment of automated or AI-supported decisions.

Pending legislative initiatives include the implementation of the Council of Europe’s AI Convention into Swiss law and the development of complementary non-legally binding measures by the end of 2026, the copyright-related legislative work following the parliamentary motion on AI misuse, and the ongoing partial revision of the Product Safety Act.

As Switzerland currently does not have any specific AI legislation, there are no general legal requirements that apply specifically to AI in terms of transparency, explainability or audits. However, transparency-related obligations may arise under existing technology-neutral laws, in particular the FADP, where personal data is involved in AI-supported data processing.

Transparency is anchored in the FADP as a processing principle and is reflected, among others, in information duties, rules on automated individual decisions and the requirement to conduct a data protection impact assessment in high-risk cases. In a communication published in May 2025, the FDPIC stated that manufacturers, providers and users of AI systems must make the purpose, functionality and data sources of AI-based processing transparent. The FDPIC also stated that users have a legal right to know whether they are speaking or corresponding with a machine and whether the data they have entered is being processed to improve self-learning programs or for other purposes. He further specified that the use of programs that enable the falsification of faces, images or voice messages of identifiable persons must always be clearly indicated and may even be unlawful in specific cases under criminal law.

The non-binding e-government standard eCH-0272 defines requirements on transparency, explainability and risks of AI systems. It is addressed mainly to federal, cantonal and communal authorities, but public-law institutions and private-sector organisations are also invited to apply the standard on a voluntary basis.

There are no specific legal requirements for human oversight or human-in-the-loop mechanisms in AI systems, as Switzerland does not have any AI-specific legislation yet. However, the FADP contains provisions on automated individual decision-making that apply whenever personal data is processed. Pursuant to Art. 21 FADP, data subjects may request that automated individual decisions concerning them be reviewed by a natural person. This applies to any decision based exclusively on automated processing that has a legal consequence for, or a considerable adverse effect on, the data subject, whether AI is involved or not.

As a matter of best practice, the Federal Administration’s strategy for the use of AI systems expects responsibilities, competencies and roles to be clearly defined throughout the entire AI system life cycle and requires administrative units to manage and monitor AI systems efficiently throughout that life cycle.

There are no AI-specific legal or regulatory requirements addressing algorithmic bias, discrimination or fairness in AI systems.

However, the principles of equality and non-discrimination are anchored in Art. 8 of the Federal Constitution of the Swiss Confederation and in several specific laws such as the Gender Equality Act (GEA) and the Disability Discrimination Act (DDA), which may also apply to AI systems where their respective scope of application is engaged. Where AI systems involve the processing of personal data, the FADP may also be relevant, through its general processing principles and rules on profiling and automated individual decisions.

With respect to future legislation, the Federal Council has confirmed that the consultation draft to implement the Council of Europe’s AI Convention shall define the necessary legal measures, particularly in the areas of transparency, data protection, non-discrimination and supervision.

In Switzerland, there are currently no AI-specific liability rules in force. However, general product liability, contractual liability, tort liability, data protection law and sector-specific liability regimes my apply to harm caused by AI systems. Liability is allocated according to the role of the party involved, the legal basis of the claim and the type of harm.

Product Liability Act (PLA)

Where an AI system is embedded in a physical product, such as a medical device, machine, vehicle or other connected product, the PLA may apply. The producer, importer or supplier may face strict liability if a defective product causes death, personal injury or qualifying property damage. A product is defective if it does not provide the safety that the public is entitled to expect. Whether and to what extent the PLA applies to standalone software or standalone AI systems remains less settled under current Swiss law.

Swiss Code of Obligations (CO)

Providers and deployers may be liable under contractual and tort liability rules:

– Contractual liability: Under Art. 97 CO, a provider may be liable for non-performance or defective performance, for example if the AI system does not meet contractually agreed functionality, performance, security or compliance requirements.

– Tort law: Under Art. 41 CO, developers, providers, deployers or users may be liable if they unlawfully cause damage intentionally or negligently, for example through incorrect configuration or integration of an AI system, use beyond the intended purpose or instructions, failure to implement appropriate safeguards or human oversight, or negligent use.

Sector-specific liability regimes

Sector-specific rules may apply in regulated industries, such as healthcare, medical devices, financial services, aviation, autonomous vehicles, and other safety-critical sectors.

Federal Act on Data Protection (FADP)

Where AI systems process personal data, the FADP applies. Failures involving unlawful processing, insufficient information, automated individual decisions, data security breaches or inadequate technical and organisational measures may trigger regulatory measures by the FDPIC and, for certain intentional violations, criminal fines of up to CHF 250’000 against responsible individuals. Civil claims for data-related harm would generally have to be based on the applicable civil-law rule, including personality protection and general liability principles.

Under the PLA, the claimant must prove the defect, the damage and the causal link between the defect and the damage. Under Art. 41 CO, the claimant must prove damage, unlawfulness, causation and fault. Under Art. 97 CO, the contractual counterparty must prove the contractual breach, damage and causation; fault is presumed, and the debtor must prove that it is not at fault.

Multiple actors may be liable depending on their respective contribution to the harm. Where several persons are liable for the same damage, joint and several liability may apply under Swiss law.

In Switzerland, there are no cybersecurity obligations that apply specifically to AI systems. However, depending on the nature of the system, the sector, its intended use and the data it processes, cybersecurity requirements under different legal frameworks may apply:

– Federal Act on Data Protection (FADP) and Ordinance on Data Protection (DPO): Whenever AI systems are used to process personal data, the cybersecurity requirements set out in Art. 8 FADP and Art. 1-6 DPO must be complied with.

– Federal Act on Information Security (ISA): The ISA applies primarily to federal authorities and certain organisations entrusted with federal public tasks. Its obligation to report cyberattacks on critical infrastructures also applies to designated operators of critical infrastructures, including private-sector operators where they fall within the statutory scope.

– Sector-specific regulations: Various sector-specific legal frameworks provide for cybersecurity obligations, including in the financial and insurance sector, healthcare, medical devices, telecommunications and automated driving.

– ICT Minimum Standards: The National Cyber Security Centre (NCSC) published ICT minimum standards for improving ICT resilience, especially for operators of critical infrastructure. Depending on the sector and applicable legal framework, these standards may be binding or serve as recommendation.

Yes. Risks related to the use of AI are generally insurable in Switzerland, depending on the type of risk and the applicable insurance terms and exclusions. They may be covered, in whole or in part, by existing insurance products such as professional liability, directors’ and officers’ liability, product liability, errors and omissions, or cyber insurance.

Cyber policies may cover AI-related incidents where the loss results from a cyberattack, data breach, business interruption or failure of IT security measures. However, they will not necessarily cover all AI-related risks, such as inaccurate AI output, discrimination, intellectual property infringement, contractual non-performance, product defects or regulatory fines.

The market trend is to address AI-related risks mainly through existing insurance lines, with more tailored underwriting depending on the insured’s sector, use case, risk profile and AI governance measures.

No. Artificial intelligence cannot be named as an inventor in a patent application in Switzerland. Under current Swiss law, only natural persons (i.e., human beings) can be listed as inventors.

This was confirmed by the Federal Administrative Court, in its decision B-2532/2024 of 4 July 2025. The Court upheld the decision of the Swiss Federal Institute of Intellectual Property (IPI), which had refused to register an AI system as the inventor in a patent application.

However, the Court clarified that a natural person may be named as inventor where they made the relevant inventive contribution, including in an AI-assisted invention process. The mere use of an AI system does not exclude human inventorship.

Swiss patent law is currently undergoing a partial revision, expected to enter into force on 1st January 2027. However, the revision does not concern AI inventorship and is therefore not expected to change this position.

In Switzerland, images or works generated solely by artificial intelligence (AI) generally do not benefit from copyright protection. The key criterion under Swiss copyright law is that a work must be an intellectual creation with individual character (Art. 2 para. 1 Federal Act on Copyright and Related Rights (CopA)). Authorship is attributed to the natural person who created the work (Art. 6 CopA).

However, works created with the assistance of AI may benefit form copyright protection if the human contribution is sufficiently creative and decisive for the final result. This may be the case, for example, where a person meaningfully selects, arranges, edits or modifies AI-generated output, or otherwise exercises creative control over the final work. In that case, authorship would be attributed to the human creator.

By contrast, purely generic prompting or the mere selection of an automatically generated output is unlikely to meet the required threshold of human intellectual creation.

When using artificial intelligence (AI) systems in the workplace in Switzerland, employers must consider employee personality rights, data protection, proportionality, transparency, non-discrimination and the limits on employee monitoring. Switzerland has not introduced AI-specific workplace regulations on AI-driven hiring, performance assessment or employee monitoring. However, existing labour, data protection and anti-discrimination rules apply. European standards, including the EU AI Act, may also be relevant for Swiss companies operating in or targeting the EU.

Employee monitoring

Employee monitoring is regulated under labour law. Art. 328 of the Swiss Code of Obligations (CO) requires employers to acknowledge and safeguard employees’ personality rights. Art. 26 of Ordinance 3 to the Employment Act (EmpO 3) prohibits monitoring or control systems that are intended solely to monitor employees’ behaviour at the workplace. Where monitoring or control systems are necessary for other reasons, they must be designed and installed in such a way as not to affect the employees’ health and ability to move.

Where employee monitoring involves personal data, employers must comply with the Federal Act on Data Protection (FADP) and the Ordinance on Data Protection (DPO), including the principles of lawfulness, transparency, proportionality and data security. Under Art. 328b CO, employers may process employee data only insofar as it concerns the employee’s suitability for the job or is necessary for performing the employment contract.

Hiring and performance assessment

AI-based hiring or performance assessment must comply with the FADP where personal data is processed, and with applicable equality and anti-discrimination rules. The FADP provisions on automated individual decisions may apply irrespective of whether AI is used. In particular, data subjects must be informed of decisions based exclusively on automated processing that have a legal consequence or a considerable adverse effect on them and may request that such decisions be reviewed by a natural person. Certain exceptions apply, including where the data subject has expressly consented.

The development and use of artificial intelligence (AI) systems raise privacy issues under the Federal Act on Data Protection (FADP) and, in cross-border contexts, potentially the EU General Data Protection Regulation (GDPR). These issues arise throughout the AI life cycle, from data collection and training to deployment and decision-making.

Key issues include lawfulness, purpose limitation, proportionality and transparency; use of personal or sensitive personal data for training; automated individual decisions; profiling; data security; data protection impact assessments for high-risk processing; anonymisation and re-identification risks; rights of data subjects under the FADP; and international data transfers.

The Federal Data Protection and Information Commissioner (FDPIC) has repeatedly confirmed that the Federal Act on Data Protection (FADP) applies directly to AI-supported data processing where personal data is involved. He further highlighted the risk that pseudonymised or anonymised data may be reassigned to a specific person with the help of AI.

The FDPIC emphasised in a recent press release, published in May 2025, that manufacturers, providers and users of AI systems must ensure transparency regarding the purpose, functionality, and data sources of AI-based processing, comply with data subjects’ rights with respect to automated individual decisions and, in the case of AI-powered conversational systems (e.g., intelligent language models), inform users that they are interacting with a machine and disclose whether input is being used to train or further develop the system.

As an enforcement example, the FDPIC’S preliminary investigation into X/Twitter’s use of public user posts to train the AI model Grok can be cited. The investigation focused on transparency and opt-out rights. After X introduced an opt-out mechanism allowing users to object to the use of their public posts for training and fine-tuning Grok, the FDPIC concluded that the FADP requirements were met and closed the preliminary investigation.

Swiss law does not contain AI-specific rules on data scraping. Its legality must be assessed under existing IP, data protection and unfair competition rules.

Copyright/IP:

Under the Federal Act on Copyright and Related Rights (CopA), copyright protection applies only to works with individual character. Raw data or factual information (e.g., basic listings or unstructured datasets) is generally not protected as such. However, scraping may raise copyright issues if protected works are copied or used, for example text images or creatively structured content. The use of copyright-protected works for AI training is currently not clearly regulated. Parliament has adopted a motion mandating the Federal Council to prepare legal amendments to better protect journalistic content and other copyright-protected works and performances against AI-related misuse; the precise legislative approach has not yet been determined.

Data protection:

Where scraped data includes personal data, the Federal Act on Data Protection (FADP) applies. Processing, including for AI training, must comply with the principles of lawfulness, transparency, proportionality, purpose limitation and data security. Consent is not mandatory under Swiss data protection law, but the controller must comply with the FADP and may need a justification where the processing infringes personality rights, for example, because it is contrary to the data subject’s express wishes or involves disclosure of sensitive personal data to third parties.

Unfair competition:

The Federal Act on Unfair Competition (UCA) may apply in specific scraping scenarios. Art. 5c UCA prohibits taking over and exploiting another person’s market-ready work product by technical reproduction processes without any reasonable effort of one’s own. Depending on the circumstances, large-scale scraping of commercially valuable datasets may therefore raise unfair competition concerns.

As of June 2026, there appear to be no Swiss court decisions directly addressing the legality of data scraping in the context of AI training. From a data protection perspective, the FDPIC’s preliminary investigation into X/Twitter’s use of public user posts to train Grok is related, although it concerned the platform operator’s use of content from its own platform rather than third-party scraping. The FDPIC closed the investigation after X introduced an opt-out mechanism allowing users to object to the use of their public posts for AI training (cf. question 13).

In Switzerland, a prohibition of data scraping in a website’s terms of use is enforceable if the terms are validly incorporated into a contractual relationship between the website operator and the scraper.

This is most likely where the scraper has accepted the terms of use, for example by registering an account, using a service subject to click-through terms, making an order, or otherwise clearly manifesting consent. In such cases, an anti-scraping clause may be contractually enforceable, and breach may give rise to legal claims. If the scraper has not accepted the terms, enforceability is less certain. Mere publication of terms of use on a website may not, by itself, be sufficient to bind a third party contractually under Swiss law. However, scraping may still be restricted under other legal grounds, including copyright, unfair competition, data protection or unauthorised access rules, depending on the circumstances.

As of June 2026, there is no central AI-specific authority responsible for supervising the use and development of AI in Switzerland. However, sectoral authorities supervise the use of AI where it falls within their existing statutory mandates. For example:

– The Swiss Financial Market Supervisory Authority (FINMA) supervises AI-related risks in financial institutions under its supervision.

– The Swiss Agency for Therapeutic Products (Swissmedic) supervises medical devices and therapeutic products, including where AI is used within regulated products or processes.

– The Federal Roads Office (FEDRO) is responsible for road traffic regulation, including automated driving.

– The Federal Data Protection and Information Commissioner (FDPIC) supervises compliance with the FADP where AI involves the processing of personal data.

These authorities use their ordinary supervisory and enforcement tools under the applicable sectorial laws. In the pending implementation of the Council of Europe AI Convention, supervision has been identified as one of the areas in which Swiss law may require further measures.

Switzerland has a long-standing reputation for excellence in research and innovation and is home to an ecosystem of multinational corporations, SMEs and start-ups. In the field of AI, Switzerland has an exceptional density of talent. According to the Stanford University’s Artificial Intelligence Index 2026, Switzerland led the global ranking for AI authors and inventors per capita in 2025, with approximately 110.5 AI authors and inventors per 100’000 inhabitants, followed closely by Singapore.

The adoption of AI is increasingly established in Switzerland and is growing across many sectors and company sizes. According to AXA’s annual labour market study, the share of SMEs incorporating AI tools rose from 22% to 34% between 2024 and 2025, and nearly half of Swiss SMEs see AI as an asset to their business, compared with 35% in 2024.

Among the sectors with significant AI adoption and/or development, the following can be highlighted:

– Healthcare and life sciences: Switzerland’s strong pharmaceutical, biotech and medtech sectors are important drivers of AI innovation, including diagnostics, clinic analytics, and medical devices.

– Finance and insurance: AI technologies are increasingly used in Switzerland’s banking and insurance sector for applications such as fraud detection, risk scoring and customer service automation. According to the EY Banking Barometer 2026, 78% of banks in Switzerland are addressing the implementation of AI, compared with 53% the previous year.

AI-driven tools are increasingly used in the Swiss legal sector by law firms and in-house legal departments, particularly for legal research, document review, drafting, translation, contract analysis, due diligence and knowledge management. Adoption is growing, and specialised AI tools for legal work are increasingly available. AI agents may become an important driver of efficiency and process optimisation in the future, but their use in legal practice remains subject to careful human review.

The main regulatory concerns are compliance with data protection law where personal data is involved, protection of attorney-client privilege and professional secrecy, confidentiality, information security, accuracy and verification of outputs, avoidance of hallucinated or incorrect legal content, and compliance with professional duties when using cloud-based or third-party AI tools.

Key challenges:

1. Data protection and professional secrecy: In addition to ensuring compliance with the Swiss Federal Act on Data Protection (FADP), lawyers must comply with professional secrecy under Art. 321 of the Swiss Criminal Code (CC), in particular when using cloud-based solutions.

2. Errors and hallucinations: When using tools based on generative AI and LLMs, human review is essential to check the quality and accuracy of the output.

3. AI literacy: With the increasing pace of technological advancements, lawyers must keep track of new AI-related risks and capabilities, enhance their skills and ensure that their personnel are trained in the use of AI.

4. Changing pricing expectations: As AI increases efficiency in legal service delivery, clients may increasingly expect predictable outputs, fixed-fee packages or value-based pricing rather than purely hourly billing. This requires lawyers to improve internal efficiency while maintaining quality, professional judgment and accountability.

5. Regulatory uncertainty: Many countries, including Switzerland, have not yet adopted any AI-specific legislation, while others, such as the EU, are still fine-tuning implementation. This rapidly changing environment is a challenge for Swiss lawyers advising on the development, deployment, and commercial use of AI technologies nationally and internationally.

Key opportunities:

1. Increased efficiency and productivity: AI can accelerate time-consuming tasks such as legal research, analysis of large datasets, document creation and summarisation, allowing lawyers to spend more time on high-value tasks.

2. Automation of routine tasks: AI tools and AI agents may automate routine tasks, such as regulatory monitoring, contract drafting, and administrative work.

3. Competitive advantage: The use of AI can help law firms offer faster and more cost-efficient services.

4. New counselling opportunities: The rapid adoption of AI and the increasingly complex regulatory landscape create growing demand for specialised advice, including on AI governance, data protection, the EU AI Act, the EU Data Act for Swiss companies within its scope, and sector-specific regulation for AI-supported products such as medical devices.

5. Innovative services: AI-powered tools allow lawyers to develop and offer new services to clients, for example specialised AI agents for specific legal or compliance use cases.

Switzerland signed the Council of Europe’s AI Convention in March 2025 but has not translated it into national law yet. A consultation draft for the necessary legislative amendments, along with proposed non-legislative measures, is expected by the end of 2026 (cf. question 2).

Clarifications on the use of copyright-protected works and performances by AI are expected in connection with the revision of the CopA, following a parliamentary motion mandating the Federal Council to prepare the necessary legislative amendments. The precise legislative approach has not yet been determined (cf. question 3).