A recent report from the Cyberspace Solarium Commission 2.0 indicates that the U.S. federal government’s cyber policy has regressed by approximately 13%, marking the first significant setback since the commission’s establishment. This decline is attributed to workforce reductions and budget cuts within key agencies, including the Cybersecurity and Infrastructure Security Agency and the State Department’s cyber diplomacy staff. The analysis highlights that nearly a quarter of the previously implemented recommendations have lost their status, suggesting a fragile progress in cybersecurity efforts. Mark Montgomery, a former Navy rear admiral, expressed concerns over the significant cuts at the Cybersecurity and Infrastructure Security Agency, stating that such workforce reductions hamper the agency’s effectiveness in addressing rising cyber threats. The report calls for the restoration of funding and personnel to bolster the nation’s cyber defense capabilities.
AI-generated code is now responsible for one in five security breaches, according to a report by Aikido. The study reveals that AI coding tools account for 24% of production code, with a particularly concerning 43% of U.S. organizations reporting serious incidents linked to AI-related flaws, compared to 20% in Europe. The report highlights that adding more security tools does not necessarily improve safety; in fact, organizations using six to nine tools saw a 90% incident rate, versus 64% for those with just one or two tools. Despite these challenges, 96% of industry professionals remain optimistic that AI will eventually produce secure and reliable code, with many expecting this to happen within the next three to five years.
A new report from SolarWinds reveals that organizations utilizing generative artificial intelligence in their IT service management significantly reduce incident resolution times. According to the 2025 ITSM Report, the average resolution time decreased from 27.42 hours to 22.55 hours after the implementation of generative AI, resulting in a time savings of nearly five hours per incident and an overall reduction of 17.8 percent. The report analyzed over 2,000 IT service management systems and more than 60,000 anonymized customer data points.
A recent report by Accenture reveals that nearly one in five employees, specifically 19 percent of office workers, admitted to entering sensitive business information into free, unsecured artificial intelligence tools. The survey, which involved 1,000 Irish office workers, highlights significant gaps in cybersecurity awareness and training, despite 65 percent of respondents receiving regular training. Jacky Fox, a senior managing director at Accenture Cybersecurity, emphasized the urgent need for businesses to address these training deficiencies and clarify cybersecurity responsibilities among staff. Concerns about AI-driven phishing and deepfake impersonation are rising, with 47 percent of employees worried about AI phishing attempts and 34 percent about identity theft.
Why do we care?
So here’s the headline: the U.S. government’s cyber defenses? Going backward. The Cyberspace Solarium Commission says we’ve lost about 13 percent of our progress — mostly because of budget and staffing cuts at agencies like CISA. Translation: less coordination, less support, more risk trickling down to the private sector.
Now add this — one in five breaches today comes from AI-generated code. Yeah, that’s right. Developers are shipping code written by AI, and the flaws are showing up in production. The kicker? The more tools companies use, the worse their outcomes — nine tools led to a 90 percent incident rate. Simplicity wins.
SolarWinds says AI can still help — they found incident resolution times dropped almost 18 percent when AI got involved. So it’s not all bad news.
But here’s the kicker from Accenture: 19 percent of employees are pasting company data into free AI tools. That’s a training and governance failure, plain and simple.
For IT service providers, the obvious takeaway is governance, which is the new firewall. Audit your AI use, cut tool sprawl, and teach clients how to stay smart about AI. Because the threat isn’t the tech — it’s the chaos that comes with using it carelessly.