Safe use of gen AI at work: how we make the most of AI while mitigating its risks

In 2026, nearly nine in 10 organizations use AI in at least one business function, and 44% are scaling AI across the organization – up six percentage points compared to last year.  As the technology becomes more widely used, and challenges such as hallucinations, output quality and data safety become even more critical, companies need clear rules and tailored safeguards to help their people use it responsibly in their everyday work.

At Allianz Trade, we give employees the tools, training and guidance they need to use AI safely and efficiently. A strong AI governance framework is central to this approach. Our framework is designed to guide us as new questions and expectations arise, keeping trust at the forefront while driving innovation and efficiency.

Solid governance is the foundation of responsible AI

Responsible AI is about making sure the technology is used safely and that this usage is accurately reported internally. At Allianz Trade, this means applying a governance framework, complying with requirements such as the EU AI Act and following responsible AI principles including:

• Diversity, non-discrimination and fairness: ensuring AI systems treat individuals and groups equitably and reducing the risk that AI reinforces or amplifies bias linked to characteristics such as gender, age, ethnicity or disability
• Transparency and explainability: informing users when they’re interacting with AI and making AI use understandable and traceable, including how systems are used and how decisions are reached
• Accountability: making sure there’s clear oversight of how AI systems are developed, deployed and monitored

Why is this important? Because while AI can optimize processes, its use also comes with risks. If companies want to benefit from its advantages, they need to set clear boundaries. This is what will allow employees to learn about and use the technology without compromising security or trust.

What lies behind the safe use of gen AI?

Every gen AI-based tool used at Allianz Trade must be reviewed by our data privacy, legal and security teams before entering a pilot phase. This gives us the opportunity to assess the use case, identify any potential risks and make sure the right controls are in place from the start. Responsibility also means keeping a human in – or over – the loop. In practice, this helps ensure that our employees remain in full control of any decision-making.

Once an AI-driven tool is deployed, we monitor it through three layers: business user feedback (with qualitative feedback from employees on whether the output was helpful), statistical sampling to ensure consistency, and technical monitoring to track model updates. For our in-house tools, we log every AI interaction – who asked what, when, and what the output was. Our AI governance framework ensures that every AI asset, whether internal or external, is clearly tagged and every new software tool, service or application is reviewed and audited before rollout. This helps us ensure any AI use is visible and aligned with our standards.

Data leakage, shadow AI and more: how do we manage gen AI-related risks?

One of the biggest risks in scaling generative AI is the mishandling of sensitive information. A key question is how to maintain confidentiality when employees use AI assistants, and how to avoid issues such as data leakage, bias and hallucinations, where the model generates incorrect or misleading results. This is why a risk-based approach to governance is essential. The goal is to set the right thresholds at every step so that employees can use AI responsibly. However, these control measures need to be implemented without hindering the pace of innovation.

Management plays a key role in making this happen. At Allianz Trade, AI literacy training is mandatory for every employee. We want our people to understand and benefit from AI while being aware of its limits and risks. We also embed controls directly into the tools we deploy. Depending on the use case, that may mean limiting what AI can do, deciding who has access or defining specific training requirements. At the same time, we need to avoid over-governing the technology. If controls become too restrictive or our tools don’t evolve at the right pace, there is the risk of “shadow AI,” where employees use alternative AI tools because our tools are either outdated or don’t serve their needs.

Setting the right guidelines to navigate the future risks of gen AI

The future of generative AI will depend on resilience as much as innovation. For companies everywhere, AI governance will only become more important as the technology becomes a bigger part of everyday business.

At Allianz Trade, we will make sure to stay ahead of the curve by strengthening the safeguards and controls that help us use AI responsibly and securely. As we explore new AI assets, including more transformative tools, we will keep asking the same key questions: “Who is responsible?”, “How is the tool controlled?” and “How does it interact safely with our systems and data?” This basis will help us continue to review and evolve our framework so it stays aligned with market developments and supports the safe, responsible use of AI over time.