How RISE Works with AI
AI (artificial intelligence) is, in addition to being one of our many research areas, a tool and an enabler used by RISE in a wide range of contexts. We strive for a responsible and curious approach that creates the greatest possible value for our customers, partners, owners, and employees.
We see significant potential for value creation with AI. We are fundamentally positive toward new technology and innovation, and our guiding principle is to lead by example. This also applies to the responsible development and use of AI – even in areas where laws, regulations, or established practices are lacking.
The number of AI technologies and applications is large, and we make continuous assessments to ensure that the benefits are proportionate to any potential risks of negative impact. Our position is that we should harness the advantages of AI in our research and operations – but never at the expense of health, the environment, or human rights.
We follow internally developed guidelines that cover several dimensions: ethical, security-related, and legal. These guidelines apply to everyone who procures, develops, modifies, or uses AI on behalf of RISE. In addition, we have specific guidelines for the use of chatbots and language models.
At RISE, there are always physical persons who hold responsibility for both the use and outcomes in cases where we employ AI. Responsibility for a result or decision can never be delegated to AI.
To provide our employees with knowledge about AI and awareness of our guidelines, we offer internal training sessions led by our experts in various aspects of AI, such as legal considerations. Employees also have access to our experts as sounding boards to help navigate opportunities and potential risks related to everything from development to procurement and use.
AI and security
Security within AI use can encompass information security, cybersecurity, and personal safety. RISE evaluates all these aspects when we develop and use AI and strives to minimize the amount of data processed in order to reduce risks, improve efficiency, and strengthen privacy protection, thereby contributing to sustainable AI use.
In research, there may be tolerance for prototypes and untested solutions, but once systems are to be tested or used in real-world conditions, the requirements for security and robustness increase significantly. Our methods are designed to account for these differing requirements.
AI and Ethics – inclusion, environment, and transparency
In RISE’s development and use of AI, we ensure that systems are equitable and inclusive. To minimize and address potential unfairness, individuals who encounter an AI system are given opportunities to report risks or negative consequences. Our ambition is also to ensure that our solutions have as little negative climate impact as possible.
As the use of AI grows, it can become difficult to determine what has been generated or modified with AI. In accordance with the AI Act, we label images that have been generated or modified using AI. While the responsibility ultimately lies with the person generating and sharing the material, RISE also works actively to adapt support and systems to make it easy to comply with requirements and guidelines.
AI and legal requirements
Several laws and regulations govern components of AI. The EU AI Act is central to RISE’s work with responsible AI, and we work purposefully with its various aspects:
- The AI Act’s risk-based framework, which classifies and ranks risks associated with different AI applications.
- The roles and responsibilities the AI Act assigns to actors within the AI value chain.
- Seven risk areas for AI highlighted in the EU’s guidelines for trustworthy AI.
Special conditions for research and development
The AI Act exempts scientific research, development, defense, and law enforcement from parts of the regulation. We evaluate and test AI for research and innovation purposes in controlled environments and actively work to identify and manage risks. For AI that will be tested under real-world conditions, the requirements increase, which is reflected in our ways of working.