Black Duck: 97% of developers now use AI coding tools — but only one-third have full governance

A Black Duck Security study published in June 2026 found that 97% of developers now use AI coding tools, but only one-third of those organizations have implemented full governance frameworks for AI-generated code. The study aggregates findings from Stack Overflow, JetBrains, GitH

The data shows AI coding has crossed the chasm into default-tool status, but the governance gap is widening: most organizations lack policies for license compliance, secret leakage, code provenance, and review of AI-generated PRs. The 97% adoption number — combined with JetBrains' January 2026 "AI Pulse" survey of 10,000+ developers — suggests the conversation in H2 2026 will shift from "should we use AI to code?" to "how do we trust, audit, and govern AI-generated code at scale?"

---

```json

[

{"title": "GitHub Copilot CLI's redesigned terminal interface hits general availability", "source_name": "GitHub Changelog", "source_url": "https://github.blog/changelog/2026-06-23-copilot-cli-new-terminal-interface-is-generally-available/", "date": "June 23, 2026", "excerpt": "GitHub announced that the redesigned Copilot CLI terminal interface — first previewed at Microsoft Build 2026 — is now generally available, giving developers a tabbed layout for working with GitHub issues, pull requests, and CI runs directly from the command line. The release makes Copilot CLI a first-class agentic shell on par with editor extensions.", "full": "The same week, GitHub also added support for organization- and enterprise-agents inside JetBrains IDEs (preview), BYOK for the Copilot desktop app (OpenAI, Azure OpenAI, Microsoft Foundry, Anthropic, LM Studio, and others), repository-level REST APIs for Code Quality findings, and Dependabot support for automatic access to GitHub-hosted registries — all pointing at a unified Copilot surface that spans IDE, CLI, app, and cloud agent."},

{"title": "AI was supposed to kill engineering jobs, but new data suggests they're the most resilient", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2026/06/24/ai-was-supposed-to-kill-engineering-jobs-but-new-data-suggests-theyre-the-most-resilient/", "date": "June 24, 2026", "excerpt": "SignalFire's State of Talent report shows that while AI dominates layoff headlines, engineers are actually gaining share of new hires across 2026. The data undercuts the 'AI replaces coders' narrative and reframes the labor story around reskilling, AI tool fluency, and rising demand for engineers who can ship AI-augmented software.", "full": "The piece complements two other June 24 TechCrunch stories: 'Companies are scrambling to stop employees from maxing out AI budgets with small tasks' (the era of 'token rationing' replacing 'tokenmaxxing') and 'AI researchers continue to leave Google for its rivals' — together sketching a 2026 labor market where AI tooling is the most valuable skill, even as headcount for non-engineering roles contracts."},

{"title": "Figma adds code layers, animations, and more AI features in a new update", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2026/06/24/figma-adds-code-layers-support-for-animations-more-ai-features-in-new-update/", "date": "June 24, 2026", "excerpt": "Figma's update adds a new code layer, support for motion and shaders, and the ability for users to create custom plug-ins for various tasks using AI. The release positions Figma deeper in the AI-assisted design-to-code loop.", "full": "The code-layer feature lets designers attach executable code snippets to any layer — turning Figma files into runnable specs that tools like Cursor and Copilot can ingest. It's a notable step in the convergence of design tools and AI coding assistants, and follows Figma's earlier launches of Figma Make and Sites for AI-driven UI generation."},

{"title": "Anthropic's Claude Tag: an always-on AI teammate learning your company from Slack", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2026/06/23/anthropics-claude-tag-is-learning-your-company-one-slack-message-at-a-time/", "date": "June 23, 2026", "excerpt": "Anthropic's new Claude Tag feature brings an always-on AI teammate into Slack. Beyond productivity, the feature is a strategic play to capture organizational context, institutional knowledge, and enterprise workflows.", "full": "Claude Tag surfaces in Slack threads, summarizes long discussions, drafts replies, and — critically for engineering teams — links answers to relevant code repos and past PRs. Combined with Anthropic's Seoul office opening (June 17–18) and its Korea enterprise wave with NAVER, Samsung SDS, LG CNS, and Nexon, the move signals Anthropic's push to make Claude the embedded AI layer across the engineering org chart."},

{"title": "OpenAI launches initiative to find and patch open-source bugs with AI", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2026/06/22/openai-launches-new-initiative-to-help-find-and-patch-open-source-bugs/", "date": "June 22, 2026", "excerpt": "OpenAI launched a new initiative to use its AI models — including a freshly released GPT-5.5-Cyber variant — to help the open-source community find and patch vulnerabilities.", "full": "The initiative is timed to a wave of AI-driven security tooling: GPT-5.5-Cyber scored 85.6% on the CyberGym benchmark (the highest single-model result to date), and Anthropic's Claude Mythos 5 is deployed under 'Project Glasswing' for cyberdefenders. For developers, this means the same AI models that write their code are increasingly auditing it — and proposing patches upstream."},

{"title": "The AI world is getting 'loopy' — swarms of agents running continuously in the background", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2026/06/22/the-ai-world-is-getting-loopy/", "date": "June 22, 2026", "excerpt": "TechCrunch reports that the next step beyond agentic AI is the 'loop': authorizing a swarm of agents to work continuously in the background, endlessly, against long-running coding and operations tasks.", "full": "The pattern is already visible in GitHub's Copilot cloud agent, OpenAI's Codex sandbox acquisition (Ona, for persistent execution environments), and Anthropic's Fable 5 / Mythos 5 long-context agents. The 'loop' framing reframes software development as a continuous async activity rather than a sequence of human-driven sessions."},

{"title": "MAI-Code-1-Flash: Microsoft's small coding model rolls out across Copilot", "source_name": "GitHub Changelog", "source_url": "https://github.blog/changelog/2026-06-18-mai-code-1-flash-available-on-more-copilot-surfaces/", "date": "June 18, 2026", "excerpt": "MAI-Code-1-Flash — Microsoft's purpose-built small coding model — is now available across additional GitHub Copilot surfaces: Copilot CLI, Copilot cloud agent, the GitHub Copilot app, and Copilot Chat.", "full": "The same week also brought updates to Copilot code review (repository-level AGENTS.md support, easier review requests on draft PRs), AI-credit usage metrics in the Copilot REST API, Copilot-authored PRs now indexed in author searches, and the deprecation of Opus 4.6 (fast). Together, the changes signal a tiered model menu: cheap in-house models for daily work, frontier models on demand."},

{"title": "SpaceX to acquire Cursor for $60B — the largest VC-backed startup deal ever", "source_name": "CNBC", "source_url": "https://www.cnbc.com/2026/06/16/spacex-spcx-cursor-acquisition-ipo.html", "date": "June 16, 2026", "excerpt": "SpaceX filed its $60 billion all-stock acquisition of Cursor (Anysphere) with the SEC on June 16. Cursor generates roughly $4 billion in annualized revenue, with $2.6 billion from enterprise accounts — making the deal the largest VC-backed startup acquisition in history.", "full": "Follow-on coverage from Yahoo Finance and The Motley Fool framed the deal as SpaceX's bid to own the AI coding stack alongside its Grok models. SpaceX later filed plans for 'Origin,' a GitHub competitor built on Cursor's developer base. The acquisition marks the first time a frontier-model company has paid a top-of-market multiple for an AI-native IDE."},

{"title": "Fable 5 shutdown: 4 open-weight coding models responded before Anthropic could restore access", "source_name": "The New Stack", "source_url": "https://thenewstack.io/fable-ban-open-weights/", "date": "June 18, 2026", "excerpt": "When the US government ordered Claude Fable 5 pulled on June 12, open-weight coding models from Cohere, Moonshot, and Zhipu became enterprises' instant second source.", "full": "CNBC's June 16 coverage noted that many of the models gaining traction are Chinese. For developers, the practical takeaway is that the closed-vs-open frontier for coding models is now a procurement decision: enterprises that bet solely on Anthropic found themselves scrambling for backups, while teams that had already integrated open-weight models (Qwen-Coder, DeepSeek-Coder-V4, GLM-Coder) kept shipping."},

{"title": "Black Duck: 97% of developers now use AI coding tools — but only one-third have full governance", "source_name": "Digital Applied", "source_url": "https://www.digitalapplied.com/blog/ai-coding-adoption-statistics-2026-50-data-points", "date": "June 21, 2026", "excerpt": "A Black Duck Security study published in June 2026 found that 97% of developers now use AI coding tools, but only one-third of those organizations have implemented full governance frameworks for AI-generated code.", "full": "The data shows AI coding has crossed the chasm into default-tool status, but the governance gap is widening: most organizations lack policies for license compliance, secret leakage, code provenance, and review of AI-generated PRs. The 97% adoption number — combined with JetBrains' January 2026 'AI Pulse' survey of 10,000+ developers — suggests the conversation in H2 2026 will shift from 'should we use AI to code?' to 'how do we trust, audit, and govern AI-generated code at scale?'"}

]

```