Responsible AI (RAI) framework: Ethical principles & guidelines
Principles and guidelines
Our Responsible AI Framework is founded on a commitment to ethical AI practices. We design and deploy AI systems that respect applicable laws and mitigate risks of harm, with the goal of contributing positively to society. Transparency is a cornerstone of our approach, with clear and understandable AI processes that foster trust and accountability. We have established robust mechanisms to support responsible AI and provide pathways to address potential harms if they arise. Data privacy and cybersecurity are paramount, with stringent measures in place to protect data.
Risk management
We evaluate AI applications to identify potential risks and impacts. We use comprehensive assessments to help us understand and mitigate potential negative consequences associated with AI systems. We implement effective strategies to minimize identified risks, including technical solutions, policy interventions, and user education, ensuring that our AI systems operate safely and ethically.
Continuous monitoring and evaluation
Continuous monitoring and evaluation are used to help our AI systems remain aligned with compliance requirements and effectiveness goals. Feedback mechanisms from users and stakeholders enable ongoing improvement, while our AI systems are designed to adapt to new information and changing environments while maintaining ethical integrity. Collaboration is encouraged to advance responsible AI practices and share best practices.
Education and awareness
Education and awareness are vital to our Responsible AI Framework. We offer comprehensive training programs for our employees to educate them about responsible AI practices, helping ensure that everyone involved is knowledgeable and prepared. We actively raise awareness about AI technologies and their implications, fostering informed, responsible and ethical use.
Documentation and reporting
Documentation and reporting are essential for transparency and accountability. We therefore maintain documentation of our AI systems, including key aspects of their design, development, and operational processes.
<ul> <li>Designed to support adherence to applicable regulations, such as the EU AI Act, GDPR, HIPAA, FDA 21 CFR Part 11</li> <li>Software development and deployment meet internal policies and procedures.</li> </ul>
Our Responsible Artificial Intelligence (RAI) Policy describes the ethical principles that anchor RAI at Thermo Fisher and colleague conduct that can help manage AI risks. This policy is intended to assist with decision making and provide colleagues with a basic understanding of appropriate ways to use, develop or deploy AI.