[webinar] Myth vs Reality: Understanding AI/ML for QA Automation - w/ Jonathan Lipps

Introduction

Hi everyone, glad you could join us for today's session, "Myth vs Reality: Understanding AI and ML in QA Automation," with our expert guest speaker Jonathan Lipps.

Introduction

Jonathan Lipps is a world-renowned test automation expert who needs no introduction, but I'll give a short one anyway. Jonathan is the architect and project lead for Appium, the popular open-source automation framework, and the author of the Appium Pro weekly newsletter. He founded Cloud Grey, a consulting firm aimed at helping global brands and companies successfully scale their mobile test automation efforts. With over 15 years as a programmer in tech startups, he holds a master's degree in philosophy and linguistics from Stanford and Oxford.

The Overarching Question: Is AI BS?

Jonathan begins with a remarkable image of a toothbrush advertised as "having artificial intelligence." This spurs the question of whether AI is fundamentally just BS—a marketing gimmick meant to appear intelligent.

What Are AI and Machine Learning?

To differentiate between hype and reality, Jonathan explains that AI is defined as "anything a computer does that seems smart." Machine learning (ML) is a field of study that gives computers the ability to learn without being explicitly programmed. He further delves into categories of ML, including supervised learning, unsupervised learning, reinforcement learning, and deep learning.

Different Categories of Machine Learning

1. Supervised Learning: Involves training a computer with labeled data.
2. Unsupervised Learning: Analyzes data that hasn't been tagged.
3. Reinforcement Learning: Focuses on teaching computers to perform tasks in specific environments through rewards.
4. Deep Learning: Deals with large amounts of complex input data.

Jonathan also covers several algorithms such as linear regression, k-means clustering, neural networks, and generative adversarial networks (GANs).

AI & Machine Learning in QA Automation

Jonathan examines how companies market AI and ML in QA tools. He categorizes them into three types:

1. AI and Marketing Only: Solutions marketed as AI but use traditional programming methods.
2. AI and ML in a Supporting Role: Utilizes ML to support features without replacing test authoring or running.
3. AI and ML as the Primary Driver of Automation: Autonomous BOTS finding bugs and writing tests.

Examples and Use Cases

He shares examples ranging from capturing user activity logs to using neural networks for visual testing such as Applitools. He emphasizes that while AI can be a buzzword, there are genuinely useful applications, especially in supporting roles like performance monitoring.

Concluding Thoughts

Concluding the webinar, Jonathan discusses the potential ROI (Return on Investment) for AI/ML in QA, the risk of job displacement, and encourages QA professionals to focus on value rather than being swayed by marketing terms.

Questions from the Audience

Participants asked about practical case studies, the future of tools like Appium, and the potential for AI to fix code issues.

Keywords

• AI
• Machine Learning
• QA Automation
• Appium
• Supervised Learning
• Unsupervised Learning
• Reinforcement Learning
• Deep Learning
• ROI
• AI in Marketing
• Visual Testing
• Neural Networks

FAQ

Q1: What are the practical applications of supervised learning in QA? A: Supervised learning can classify new data instances based on trained models. For example, predicting software bugs based on historical bug data.

Q2: What differentiates AI from ML? A: AI broadly refers to anything a computer does that seems smart, whereas ML refers to algorithms that allow computers to learn from data without being explicitly programmed.

Q3: How can ML be integrated into existing QA automation tools like Appium? A: Through plugins and an ecosystem of easily integrated tools; examples include visual element recognition and OCR plugins.

Q4: Does the use of AI/ML in QA automation mean job displacement for QA professionals? A: The goal is more about enhancing human abilities by taking over repetitive tasks, but there will always be a need for human oversight and creativity.

Q5: Are there AI solutions that can fix issues in application code automatically? A: Static analysis tools exist that can suggest fixes, though fully automated code fixing is complex and not common in current QA automation tools.