Artificial intelligence / AI | Lawyer and legal services

Applications incorporating artificial intelligence (AI) have become commonplace in our lives. More companies are focusing on the development of AI solutions. Although artificial intelligence and digital solutions are intended to contribute to increased innovation and competitiveness, efficiency, and security, there are many legal pitfalls.

In a world where technology is evolving faster than ever, it's important to stay updated on regulations. Our specialist lawyers understand the complexities of the EU's AI Regulation and offer expertise to navigate this landscape, identifying risks so that you are in control of these when adopting technology or developing new digital systems.

Advokatfirmaet Halvorsen & Co assists developers, suppliers, and businesses that use or wish to implement artificial intelligence and other digital technologies. We have a strong understanding of the risk elements you may be exposed to, which will benefit you, whether it concerns assistance with contract preparation, negotiations for the sale or purchase of AI applications, or the implementation of AI products within your business.

The AI Regulation has been adopted

The European Parliament adopted the Union's new on 13 March 2024 Artificial Intelligence Act. The law must undergo formal approval by the EU Council of Ministers before it can enter into force in the EU. The formal adoption of the AI Act is expected to take place shortly, and the law will then enter into force 20 days after its publication in the Official Journal of the European Union.

The government had a working group in autumn 2023 that assessed EEA relevance, and this group concluded that the legal act is EEA-relevant. When the regulation is formally adopted, it will have to be directly incorporated into Norwegian law. Because Norway has taken an active approach and contributed to EU legislation, we expect Norway to implement this EU legislation into Norwegian law relatively quickly after the law has entered into force in the EU.

AI Act

The regulation is based on a risk-based approach, with classification into risk categories. This means that the higher the risk posed by the use of artificial intelligence, the stricter the use is regulated. The proposed rules therefore primarily regulate «high-risk» uses of AI, which pose a significant potential threat to society and individuals. The aim is to make the EU the best in the world in developing and using safe, reliable and human-centric artificial intelligence, because trust in AI systems is important and necessary for the social and economic potential of artificial intelligence to be fully realised.

The AI Act divides the use of artificial intelligence into four categories:

1 Unacceptable risk – prohibition

This includes AI systems that are considered unacceptable because they violate fundamental
values in the EU, for example breaches of fundamental rights. The ban includes, among other things:

AI systems that have significant potential to manipulate individuals through subliminal techniques beyond their consciousness, thereby influencing their behaviour in a way that is likely to cause psychological or physical harm
Exploiting the vulnerabilities of a group due to age or mental or physical impairment to influence their behaviour in a way that is likely to cause mental or physical harm
AI-based systems used by public authorities to assess the trustworthiness of individuals based on social behaviour in various contexts, or by using known or predicted personality traits, and which may lead to detrimental or disadvantageous treatment of that person or group of persons in contexts unrelated to the contexts for which the data was originally generated or collected, detrimental or disadvantageous treatment of certain natural persons or entire groups, which is disproportionate to the seriousness of their social behaviour
the use of real-time biometric identification systems located in publicly accessible spaces (mass surveillance) for law enforcement purposes.

2 High risk

This category will include AI systems that entail a high risk to people's health, safety
or fundamental rights. These AI systems will be lawful provided that:

The systems are subject to a risk management system.
The data the systems are built on is of high quality to ensure that the results are impartial and do not discriminate.
Detailed documentation is provided on how the AI systems function, enabling authorities to assess whether regulations have been complied with.
Suppliers share information with users to help them understand and use AI systems correctly.
There must be human oversight in both the design and implementation of artificial intelligence.

3 Limited risk

Certain AI systems are subject to certain transparency obligations. This applies to systems intended to interact with natural persons. These systems must be designed in such a way that the person is informed that they are interacting with an AI system, if this is not obvious from the context of use. This duty to inform does not apply to AI systems regulated by law that are intended to detect, locate, identify, or prosecute criminal offences, unless they concern systems that are publicly available and established for reporting criminal offences.

4 Minimal risk

The majority of AI systems in use represent minimal risk to individuals and society as a whole. AI systems not covered by the first three categories are considered to be minimal-risk systems and are therefore not subject to specific obligations.

It is assumed that the main body of AI systems will fall into categories 3 and 4.

Oversight/sanctions

A national supervisory authority for artificial intelligence shall be established to oversee compliance with the law. The supervisory authority shall be empowered to impose fines for breaches of the regulations.

The size of the fines depends on which articles have been breached. The most serious infringements may result in fines of up to 1.51% of global turnover, up to a maximum of €35 million, or 71% of global turnover, depending on the nature of the infringement and the size of the company.

Other promises

Beyond AI law, there are several other legislative regulations in the pipeline within the digital sector, including
AI Liability Directive which will regulate non-contractual civil liability for damage caused by
AI system, Data Governance Act, Data Act, Cyber Resilience Act, Digital Markets Act.

How can I help?

We can help you ensure that you comply with regulations, in assessments about whether you are entitled to use and
data deletion, holding it up against related areas such as privacy (GDPR), copyright, contractual
rights and obligations, confidentiality/trade secrets, security legislation etc.

We can also assist with

privacy assessments, including
- if the business has sufficient grounds to process the personal data for the
  planned purposes in a lawful manner,
- which personal data is processed,
- What privacy features does the tool have and for what purposes will the personal data
  dealt with.
Data Protection Impact Assessments
design and negotiation of contracts for the sale and purchase of AI applications
risk assessments
internal guidelines for the use of artificial intelligence
employment law requirements during implementation, including
- Requirements for consultation under the Working Environment Act and any collective agreement
- Requirements for employee training and guidance
- Use of AI in redundancy cases and downsizing

Artificial intelligence (Artificial Intelligence)