Fast Company raises a question it does not answer: whether companies will let AI change live applications, following the automation of monitoring and observability work. Independent commentary from Splunk and Swarmia repeats the instability finding, linking AI adoption to more change failures and longer recovery times. DORA frames AI as an amplifier of existing strengths and weaknesses, not a standalone fix. The instability pattern also appeared in DORA's 2024 report, so the 2025 edition sharpens an existing finding rather than introducing a new one. Beyond the 2025 report's summary, the available commentary is secondary, and no details on whether any specific company has granted AI control over production systems are public.
The story matters because about 90% of developers now use some AI assistance, according to DORA. Attorneys advising software companies or their customers should watch for contract terms covering uptime, incident response, and liability for failures caused by AI-generated changes. DORA's list of factors that determine whether AI speed produces better outcomes, including small batches, version control, platform engineering, and automated testing, offers a practical checklist for due diligence. Clients that have invested heavily in AI coding tools and now want to automate operations should be asked whether those foundations are in place.