artificial intelligence

Artificial intelligence is transforming software testing on two levels: First, an increasing number of applications are emerging whose behavior is shaped by machine learning, training data, and models. Second, testing teams are using AI themselves to analyze data more quickly, develop test ideas, design test cases, or support test automation. 

This does not replace software testing – but it is changing it significantly.

Traditional software generally follows clearly defined rules: a specific input leads to an expected output. AI-based systems work differently. They deliver results based on data, probabilities, and learned patterns. This makes them powerful, but also harder to predict.

For testers, this means: It is no longer enough to simply check functions. Data quality, model behavior, bias, robustness, explainability, and risks must also become part of the test strategy. 

At the same time, Generative AI opens up new possibilities in the testing process. With the right prompts, requirements can be analyzed, test conditions derived, test cases formulated, test data ideas developed, or scripts prepared for test automation. What matters is not just that AI is used, but how controlled, traceable, and responsible its use is. 

This is precisely why topics such as Explainable AI, test data governance, and the EU AI Act are gaining importance. Anyone who wants to use AI professionally in software testing needs to understand the risks associated with AI systems, how test results can be evaluated, and where human review remains indispensable. 

At trendig, we therefore look beyond just tools. We focus on competencies: What do you need to understand to test AI-based systems? And what skills do you need to effectively use Generative AI in software testing?

AI Testing means testing AI-based systems in such a way that their behavior, risks, and quality can be assessed in a transparent manner. The focus is not on using AI as a tool in the testing process, but on testing the systems that themselves use artificial intelligence. 

This is particularly relevant for applications that use machine learning. Such systems make decisions or generate results not only based on fixed rules, but also on data, models, and learned patterns. This raises new questions for software testing: 

• How good are the training- and test data? 
• Does the model react in a robust manner to new or unexpected inputs? 
• Are there distortions or biases in the results? 
• Are decisions explainable and traceable? 
• Does the quality of the model change over time? 

AI Testing therefore expands traditional software testing to include data- and model-related perspectives.

In addition to functional testing, aspects such as data quality, model metrics, test oracles, drift, fairness, robustness, transparency, and Explainable AI are becoming increasingly important. Testers must understand how AI systems are created, how they are trained, and what risks may arise in the various phases of the AI lifecycle.

This is exactly where the ISTQB® Certified Tester AI Testing (CT-AI) v2.0 training comes in. It teaches the fundamentals of testing AI-based systems in a structured way—from the data used and the model to the behavior of the entire system. As such, CT-AI is aimed at anyone who not only wants to use artificial intelligence in software testingbut also wants to evaluate it professionally.

AI Testing helps modern testing teams avoid treating AI systems as a black box. It lays the foundation for making quality visible, identifying risks early, and making informed decisions about the use of AI-based software.

GenAI Testing refers to the targeted, controlled, and traceable use of generative AI in software testing. The focus here is not on testing an AI-based system, but rather on how testers can effectively use tools such as large language models for their testing tasks.

Generative AI can support test teams in many areas: analyzing requirements, deriving test ideas, formulating test cases, generating test data, preparing automation scripts, or summarizing test results. This makes AI a tool that can accelerate and enhance daily testing work.

However, it is crucial to note that GenAI does not provide ready-made answers. Responses may be incomplete, incorrect, distorted, or out of context. Therefore, the use of GenAI in software testing requires clear prompts, expert review, and an awareness of risks such as hallucinations, bias, data protection, security, and the unintended disclosure of sensitive information.

This is where prompt engineering becomes a key testing skill. Anyone who wants to use Generative AI in software testing must clearly formulate tasks, provide context in a meaningful way, critically evaluate results, and know when a human decision remains indispensable. GenAI can make good suggestions—but the responsibility for quality remains with the team.

The ISTQB® Certified Tester – Testing with Generative AI training course addresses exactly this point. It demonstrates how Generative AI can support the testing process, what opportunities arise for analysis, design, automation, and reporting, and what limitations testers need to be aware of.

This makes GenAI Testing more than just tool usage: it’s about effectively integrating AI into existing testing processes without relinquishing control, traceability, and quality.

Not all AI in software testing is the same. That’s why the most important question to ask at the outset is: Do you want to test AI-based systems—or do you want to use generative AI in the testing process?

If you want to understand how machine learning systems work, how training and test data are evaluated, and how to systematically test AI models, ISTQB® Certified Tester AI Testing (CT-AI) v2.0 is the right choice for you.

If you want to use generative AI to support test analysis, test design, test automation, or reporting, ISTQB® Certified Tester – Testing with Generative AI is a better fit for your goals.

Both courses complement each other. CT-AI strengthens your understanding of the quality of AI-based systems. CT-GenAI shows you how to use AI as a tool in software testing productively and responsibly.

Anyone who wants to professionally use or test artificial intelligence in software testing needs a solid understanding of the fundamentals behind AI systems. Three topics are particularly important here: machine learning, explainable AI, and test data governance.

Machine learning is the technical foundation of many AI-based applications. Such systems are not programmed in the traditional sense, but rather learn patterns from data. This is precisely why their quality depends heavily on the data used to train, validate, and test them. For software testing, this means that data is not merely preparatory material; it is a central component of quality assurance.

If training data is incomplete, unbalanced, or biased, the model may produce incorrect or unfair results. If the actual input data changes during operation, model quality may decline. Therefore, testers must not only verify whether a system works, but also whether data quality, model behavior, and the operational context align.

This is where test data governance comes into play. It ensures that test data is used in a traceable, appropriate, protected, and compliant manner. This is particularly crucial for AI systems: What data is permitted to be used? Is it representative? Does it contain sensitive information? Is it current enough? And can one later still trace why a specific test result was produced?

Another key topic is Explainable AI. AI systems can be powerful, but their decisions are not always easy to understand. For software testing, it is therefore not enough to simply accept a result. Teams must understand why a system arrives at a result, which influencing factors are relevant, and where the limits of explainability lie.

Regulatory requirements are also becoming more important. The EU AI Act places greater emphasis on risks, transparency, and responsibilities when using AI. For companies, this means that anyone who develops, implements, or tests AI must consider quality not only from a technical perspective, but also from organizational and legal standpoints.

This is precisely why these topics are an integral part of modern software testing with artificial intelligence. It is not just about tools or automation. It is about making AI systems controllable, explainable, and usable in a responsible manner.

AI is also transforming test automation. Not because it simply replaces existing automation, but because it creates new possibilities: in the analysis, design, implementation, and maintenance of automated tests.

Generative AI, in particular, can help test teams move more quickly from an idea to a first concrete result. Test ideas can emerge from requirements. Test cases can be formulated from test ideas. Suggestions for automation scripts, test data, or test conditions can be derived from test cases. This saves time—especially when teams have to handle many variants, recurring tasks, or complex domain contexts.

Nevertheless, test automation with AI is not merely a question of tools. Good results only emerge when testers know what they want to achieve, what inputs they provide to the AI, and how they evaluate the results. A generated test case is not yet a good test case. A generated script is not yet stable automation. And a plausible-sounding answer is not yet a reliable test result.

That is why AI-supported test automation needs clear guidelines:

• Which tasks is GenAI allowed to support?
• Which results must be reviewed by subject matter experts?
• What data is allowed to be entered into AI tools?
• How are prompts, decisions, and results documented?
• Who is responsible for quality and approval?

AI becomes particularly exciting in areas where test automation has traditionally required a lot of manual groundwork: structuring requirements, identifying test gaps, generating test cases, refactoring test scripts, or explaining failed tests. In such situations, AI can accelerate the testing process and give testers more time for analysis, evaluation, and quality decisions.

This means that test automation with AI is not a sure thing, but rather a matter of competence. Teams must learn to use AI in a targeted manner, critically review results, and build automation in a way that remains maintainable, traceable, and controllable.

This is precisely where the two perspectives of the training courses converge: ISTQB® Certified Tester – Testing with Generative AI demonstrates how GenAI can support the testing process and automation. ISTQB® Certified Tester AI Testing (CT-AI) v2.0 helps participants understand how AI-based systems themselves are tested and evaluated. Together, they provide the foundation for teams that want to not only experiment with AI in software testing but also use it professionally.

Artificial intelligence in software testing requires more than just curiosity about new tools. It requires in-depth knowledge, practical application, and a clear understanding of where AI can help – and where testers need to pay close attention. That’s why we offer two ISTQB® training courses that cover different perspectives on AI and software testing.

Are you interested in testing AI-based systems and understanding how to assess quality in machine learning applications? Then the ISTQB® Certified Tester AI Testing (CT-AI) v2.0 is the right course for you.

In this training, you’ll learn how AI systems differ from traditional software, the role data plays in testing, and how to assess risks related to model behavior, bias, robustness, transparency, and Explainable AI. You’ll explore testing input data, models, and ML-based systems—and develop an understanding of how quality in AI systems becomes visible and verifiable.

CT-AI is right for you if you:

• are testing AI-based systems or plan to test them in the future
• want to understand how machine learning is changing quality assurance
• want to systematically evaluate data quality, model behavior, and risks
• want to prepare for the ISTQB® CT-AI certification

In short:

CT-AI helps you not only view AI systems as a black box, but also test them in a structured and professional manner.

Are you interested in using generative AI in the testing process and want to know how to use AI tools effectively, in a controlled manner, and responsibly? Then ISTQB® Certified Tester – Testing with Generative AI is the perfect place to start.

In this training, you’ll learn how GenAI can support test analysis, test design, test automation, and reporting. You’ll explore prompt engineering, typical use cases in software testing, and the limitations of generative AI. This also includes risks such as hallucinations, bias, data protection, security, and handling sensitive information.

CT-GenAI is right for you if you:

• Want to use GenAI in software testing in a practical way
• Want to formulate prompts for test tasks more effectively
• Want to use AI support for test cases, test data ideas, or automation
• Want to understand how to critically evaluate GenAI results
• Want to prepare for the ISTQB® CT-GenAI certification

In short:

CT-GenAI helps you not only try out Generative AI, but also integrate it into your testing work in a productive and controlled manner.

Both courses complement each other. ISTQB® CT-AI shows you how to test AI-based systems. ISTQB® CT-GenAIshows you how to use AI in the testing process.

Together, they prepare you for a software testing environment where artificial intelligence is no longer a niche topic but an integral part of daily quality assurance work.