Generative Artificial Intelligence (Gen-AI) is beginning to reshape how we live and work — and it’s happening fast. Recent breakthroughs in large language models and other advanced technologies have only further accelerated this progress. We now have AI systems capable of processing unstructured content in ways we’ve never seen before, revealing opportunities to streamline workflows, enhance decision-making, and uncover untapped potential across various business functions.

However, this wave of opportunity also introduces unique challenges and risks that require thoughtful management. As companies eagerly embrace this technology, it’s important we do so responsibly and intentionally. We must ensure that our employees, partners, and customers are ready to use these tools in safe and ethical ways. This is why we developed Onna's AI Principles.

Our Principles

Recognizing the potential of AI and acknowledging the risks associated with it, we began outlining our AI ethics framework in 2020 to shape the development and use of responsible AI at Onna. As a data management platform, we combine data processing with native AI capabilities, like Natural Language Processing (NLP) and Machine Learning (ML). We then index the data to make it searchable and usable for downstream use cases. Our goal in creating this framework was to ensure that organizations collaborating with us have a clear understanding of our approach to AI.

Today, these six principles serve as a roadmap for the use of AI at Onna. To further our commitment to transparency, we believe in sharing these principles widely and empowering our customers, partners, and the broader community to engage with our technology in an informed and conscious way.

1. AI at Onna is Trustworthy

We pledge to uphold trust by actively protecting all internal, end-user, and customer data. Our commitment includes empowering our customers with full control over their data and processes. We firmly believe that the choice to apply AI to data should lie entirely in our customers' hands.

We, at Onna, will not employ customer data to train ML models without obtaining clear permission from the customer. For instance, should a customer wish to create a customized ML model based on their data, we would proceed in training the model only with their explicit consent.

2. AI at Onna is Ethical

We recognize the inherent bias in many AI systems, and commit to assessing the ethical implications while prohibiting harmful or discriminatory uses. Our dedication lies in sourcing high-quality AI models from reliable vendors, ensuring the accuracy, reliability, and safety of our AI solutions.

3. AI at Onna is Efficient

At Onna, we will use AI to boost work efficiency and enhance productivity while reinforcing human accountability for actions and decisions stemming from AI-powered processes. AI systems not only add value by automating manual tasks but also pave the way for autonomous operations.

A system that acts autonomously can operate without human intervention, producing significant productivity gains as it augments human work and liberates individuals for more personalized tasks. Our procedures echo our belief that those who design, deploy, and use the AI system must take ownership of its actions and decisions.

4. AI at Onna is Useful

In ensuring the proper use of AI, we need to look beyond just the system; we must also scrutinize how the system is being utilized. Our responsibility extends to ensuring that both the input to and output from AI are used responsibly. This includes understanding the limitations of AI, preventing illegal or unsafe behaviors, steering clear of automated AI decision-making that could cause harm without human review, and maintaining compliance with all applicable laws and regulations.

5. AI at Onna is Regulated

Onna's Ethics Committee regulates the use of AI systems and promotes best practices within Onna. This includes the use of AI within our products and services. Our stringent controls and policies, which govern our internal IT systems, also apply to our AI systems.

By implementing strong security protocols and adhering to data-security best practices, we ensure the protection of both employee and customer data. At Onna, we continually protect both individual and enterprise data, firmly upholding our commitment to privacy, security, and applicable regulations.

6. AI at Onna is Transparent

We are committed to transparency regarding our AI practices, technology, vendors, and data usage. We document and maintain the approved internal AI systems and usage procedures in Onnapedia, our internal Confluence site.

If customers choose to use machine learning with Onna, we will endeavor to provide users with a clear understanding of how our AI systems operate and the reasoning behind the AI output. This approach ensures that users are provided with appropriate context whenever possible.

Moving forward with AI

While the AI hype in the enterprise isn't new, this time feels different. As companies race to announce the latest GPT plug-in, we can expect to hear about many advancements made possible by AI, as well as the inevitable failures that come with experimentation.

Our AI Principles underscore our commitment to using AI in a way that best serves our team, partners, and customers. We believe these principles, upheld and monitored by our Ethics Committee, form a solid foundation for confidently reaping the benefits of AI without compromising the integrity of proprietary data or operations. As AI technologies advance and our knowledge of them expands, we’ll adapt our approach over time. We hope that sharing these AI principles has been helpful, and we look forward to navigating the future of AI together.

Check out these four steps to help your organization get started with generative AI ethically and responsibly.