Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsiTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
AI can help QA teams turn requirements and recorded browser journeys into draft test cases and scripts, but it cannot establish on its own that those tests are correct. A reliable workflow uses AI for suggestions and editing, then relies on testers to verify expected results, review the code, run it in the target environment, and maintain it.
What “AI in testing” means
The phrase describes two related but different activities:
- Using AI to support testing: applying generative AI to tasks such as reviewing acceptance criteria, drafting test cases or scripts, exploring defect patterns, proposing synthetic data, and preparing documentation. These are candidate outputs, not evidence that a test is valid. ISTQB lists these as potential uses across the test process in its CT-GenAI syllabus.
- Testing a system that contains AI: evaluating the AI component and the surrounding system, including its data, model, and development process. This calls for software-testing practices selected to suit the risks of the AI system, rather than simply using a chatbot to write tests.
This article focuses on the first activity. The second is a distinct discipline, covered below.
Recommended Free Tools
How to move from manual browser checks to an AI-assisted workflow
Treat the process as a loop: establish what should happen, capture or draft a test, use AI to improve the draft, review and run it, then maintain the resulting test. Microsoft documents a Power Platform example in which Playwright records browser interactions and an AI assistant rewrites the recording to follow toolkit conventions. The guide recommends reviewing and committing the generated test; it does not make review optional. See Microsoft’s AI-assisted testing overview.
- Start with a trustworthy test basis. Use requirements, acceptance criteria, existing tests, or a user journey that the team has observed. Ask AI to identify ambiguity and suggest test objectives. Check its interpretation against the product rules before treating any suggestion as a requirement.
- Record a representative browser journey. For an end-to-end browser check, use Playwright codegen to capture a happy path. Microsoft’s example then uses an assistant to rewrite that recording for the project’s Power Platform testing conventions. The recording is a starting point, not a complete test plan. GitHub Docs also describes creating end-to-end tests for a webpage.
- Ask for useful variations. Have AI propose edge cases and data variants relevant to the journey—for example, invalid input or a boundary value where the product rules define one. For each case, decide what result should count as correct. Discard suggestions that do not apply or whose expected outcome cannot be justified.
- Review the draft as test code. Inspect locators, assertions, setup and cleanup, data isolation, and alignment with the team’s framework conventions. A script that runs successfully can still assert the wrong outcome or encode a mistaken assumption.
- Run, investigate, and commit deliberately. Execute the reviewed test in its intended environment. Inspect failures to distinguish product defects from test defects, stale assumptions, and nondeterministic behavior. Preserve reproducible evidence for decisions, and commit only after the test’s purpose and behavior are clear.
- Maintain the test as the product changes. Revisit its assumptions, test data, and expected outcomes when requirements or interfaces change. AI can help draft an update, but a person still needs to decide whether the revised test reflects the current product behavior.
Where AI can help—and where judgment remains essential
AI can reduce the effort of producing and adapting test artifacts, especially when a tester can supply clear context and established project conventions. ISTQB describes possible support across acceptance-criteria review, test-case and script generation, potential-defect analysis, defect-pattern analysis, synthetic test data, and documentation. The practical value is in accelerating drafts and analysis, not transferring responsibility for quality decisions.
The key constraint is the test oracle: a credible way to determine the expected result. ISO’s report on testing AI-based systems identifies the difficulty of determining expected results—and therefore whether tests passed or failed—as a central challenge. That problem also matters when AI is helping author ordinary software tests: a fluent script can encode a weak assertion, a mistaken requirement, or an unjustified expected result. Review cannot make an unknown expected outcome known; the team must establish the product rule or otherwise define a defensible oracle.
Generative AI also brings risks such as hallucinations, bias, security, and privacy, which ISTQB’s CT-GenAI v1.1 update announcement highlights. Avoid sending sensitive source code, customer information, credentials, or test data to an AI service unless its approved data-handling terms and your organization’s policies permit it. Check generated claims and code rather than relying on confident wording.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchChoosing between manual checks, scripted automation, and AI-assisted authoring
These approaches are not mutually exclusive. A manual check may be appropriate for exploratory judgment; conventional automation may be suitable for stable, repeatable behavior; AI assistance may help draft or adapt tests. Choose by the work’s risk and evidence needs, rather than assuming that more AI means better testing.
| Decision factor | Questions to ask | What it means for the approach |
|---|---|---|
| Risk and impact | How serious would an undetected failure be, and how much confidence is needed? | Higher-impact behavior calls for stronger review and evidence, regardless of how the test was authored. ISO/IEC TS 42119-2:2025 uses a risk-based approach to selecting practices for AI systems and components. |
| Expected-result oracle | Can the team state a credible expected result? | If not, first resolve the ambiguity or define how correctness will be assessed. AI-generated cases do not solve an unknown oracle. |
| Human review | Does the draft need product, domain, security, or accessibility judgment? | Keep the relevant reviewer in the loop where the answer depends on expertise or policy, rather than accepting generated output as proof. |
| Framework fit | Does the test follow the project’s conventions and use reliable setup, locators, and cleanup? | AI may help adapt a draft, but reviewers should verify compatibility and maintainability against the actual framework. |
| Reproducibility and maintenance | Can another tester reproduce the result, and will the test remain understandable as the product changes? | Prefer clear assertions, controlled data, and recorded evidence over opaque or fragile generated scripts. |
How testing AI-based systems differs
When the product itself includes an AI component, testing it is not the same as using generative AI to write conventional browser tests. ISO/IEC TS 42119-2:2025, edition 1, published in November 2025, provides requirements and guidance for applying the ISO/IEC/IEEE 29119 software-testing series to AI systems and their components. Its catalog description says the practices are selected through a risk-based approach. ISO’s preview explains that the applicable practices include manual and automated testing, scripted and unscripted testing, and functional and non-functional testing. See the ISO catalog entry and ISO preview.
For a learning path focused on testing AI systems, ISTQB’s Certified Tester AI Testing (CT-AI) v2.0 covers input-data testing, model testing, and machine-learning development testing. The page recommends accredited training and also identifies self-study as an option.
Rank #4
ISO/IEC TR 29119-11:2020 discusses testing AI-based systems and the test-oracle challenge. The ISO catalog entry indicates that the report is under review, so check the catalog for current status before relying on it as current guidance.
Standards and guidance for test-automation tools
IEEE 3407-2025 is an active standard for end-to-end software-testing automation tools. The IEEE Standards Association page states a publication date of April 24, 2026, and ANSI approval on August 26, 2026. It concerns automation tools; it should not be confused with a standard that validates AI-generated tests or removes the need for review.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
