AI in the software development lifecycle: failure modes and fixes @Udacity
AI in the software development lifecycle: failure modes and fixes  @Udacity
Uploaded October 2026 | Updated November 2026, 3 days ago
AI is changing how software gets built, planned, tested and maintained. This session shows where it helps, where it fails, and how to keep your team in control as it evolves.

Dr. Dan Grahn, seasoned AI researcher, cybersecurity expert, walks through the software development lifecycle (SDLC) and the failure modes AI introduces at each phase. He then gives four cornerstones and five skills that keep teams stable while the tools keep changing.

In this webinar you'll learn:
• Why AI changes how the SDLC happens, not what it is
• Failure modes by phase: scope creep in planning, package hijacking in implementation, coverage that overstates test quality, and error masking in maintenance
• The four cornerstones: keep a human in the lead, work up in abstraction, apply discipline at scale, and take care of yourself
• The five essential AI skills: understand, explore, direct, evaluate and use responsibly
• How spec-driven development, mutation testing and security scanning fit into an AI-assisted workflow

Chapters
0:00 Welcome and introductions
6:22 "Work that's too fine, too early"
7:28 What the SDLC is, and what AI changes
11:40 One app, built and retired in two hours
13:57 Jobs, adoption and why full autonomy isn't ready
18:34 AI failure modes across the SDLC
26:33 Four cornerstones for long-term stability
34:58 Five essential AI skills
39:02 Where this leaves software teams
42:15 Upcoming courses
43:27 Live Q&A: requirements, spec-driven development, security, tokens, careers

Keep learning
• AI Engineering with Claude Nanodegree: udacity.com/course/ai-engineering-with-claude--nd7426
• AI-Powered Software Engineer Nanodegree: udacity.com/course/ai-powered-software-engineer--nd770

Presented by Ascendient Learning and Udacity, part of Accenture LearnVantage.

#AI #SDLC #SoftwareDevelopment #AIengineering #AIcoding
AI in the software development lifecycle: failure modes and fixesWhat It Takes to Become an AI or ML Engineer | Full WebinarWhy You Keep Logging In: Zero Trust Security Explained SimplyMost of you should not install it | OpenClaw Explained #aiagentsBuild an AI Coding Agent That Works While You Sleep (Live Demo)The Moment Our Startup Became Legit 🚀Marty Cagan on the Future of Product Management & AIOur Agentic AI Nanodegree is now FREE to startVibe Engineer, what do they do? #vibecoding #softwarearchitect #softwaredevelopmentStop giving the same AI prompts over and over 🛑How She Built a #1 App in 2 Weeks | Tech Career AdviceWhy is CLAUDE.md important? #claudecode #vibecoding #podcast
Udacity |

AI in the software development lifecycle: failure modes and fixes

SHARE TO X SHARE TO REDDIT SHARE TO FACEBOOK WALLPAPER