Artificial intelligence is adding new layers of complexity to software development security, but a Dark Reading roundtable of security executives suggests the remedy is not more tools—it is fixing the fundamentals. The central concern is identity: who can access code, pipelines, and infrastructure, and whether those permissions are visible and properly governed. Without that visibility, AI can magnify existing weaknesses rather than improve the process.
The roundtable participants emphasized that collaboration among security, development, and operations teams is essential to closing these gaps. Better identity visibility allows organizations to spot and remediate risky access before AI is introduced into the workflow. The group also noted a practical benefit: when identity controls are tightened, testing costs drop because security teams spend less time chasing misconfigured permissions and can focus on higher-value checks.
The source is a single roundtable report, so there are no dissenting views to compare. Still, the message is consistent: AI adoption in the SDLC should wait until identity gaps are closed, and doing so can deliver both security and cost advantages.