The AI Workforce Consortium today reported that the share of cybersecurity job postings across the Group of Seven economies that ask for artificial intelligence skills doubled over the year, and junior hiring has not moved with them.

This was the first in a new series of spotlight reports, this one on the impact of AI agents on cybersecurity. Job demand across the G7 grew 9.5% in the six months through March, measured against the same period a year earlier. The preceding half-year came in at 6.8%.

AI skills turned up in 28.5% of those postings, against 14.2% a year before. The most recent month in the data ran at 29.7%. All of it covers G7 markets only.

Sort the same postings by seniority and the picture changes. Senior-titled roles grew 65% in the six months through March. Growth in junior-titled roles was 5.9%. In the half-year before that they had shrunk slightly. The consortium calls this the experience paradox. Employers are asking for senior skills in jobs that are nominally entry level.

Cisco Systems Inc., which founded the consortium, asked security leaders what they cannot find in entry-level candidates. Hands-on experience with AI agents was the top gap, at 49%. Technical cybersecurity depth was 48%, human skills 45%. Four percentage points separate the top and bottom of that list, so the three are presented as a cluster rather than a ranking.

The same survey asked about budgets. Leaders picked up to three investment priorities for the next two years, and AI-powered defense was chosen most often, by 36%. People came fourth on 25%, level with threat intelligence.

Anthropic PBC disclosed a cyber-espionage campaign in November 2025 in which its Claude model ran 80% to 90% of the operation under limited human direction. In early July, Sysdig Inc. documented a fully autonomous ransomware operation.

Two weeks after that came an intrusion at Hugging Face Inc. run end to end by an autonomous agent system. OpenAI Group PBC later said said two of its models were involved, one of them unreleased. Cyber refusals had been reduced on both for an internal evaluation.

The Five Eyes cybersecurity agencies issued a joint warning on frontier models on June 22. “The timeline is not years, it is months,” they wrote.

Roles are being rebuilt around that shift. The tier-one security operations center analyst is the clearest case. That job was manual triage. It is now closer to supervising agents and checking their work. Digital forensics analyst was the fastest-growing title in the posting analysis, up 453%, though from a base of roughly 150 to 200 AI-skilled forensics listings per half-year.

“One of the roles that is getting reshaped the fastest is the tier 1 SOC analyst,” Omar Santos, a distinguished engineer at Cisco, said during a consortium webinar in May. Triage, threat intelligence checks and data correlation are “where agentic systems are now generally effective,” he said. “Everybody is an orchestrator now.”

Ethical reasoning showed up in 533% more cybersecurity listings than a year earlier. Systems thinking was up 251%. Stakeholder engagement, 125%. The bases are small in all three cases, and the authors treat the numbers as direction rather than a measure of market size.

IBM Corp.’s Cost of a Data Breach Report 2025 found 63% of breached organizations either had no AI governance policy or were still writing one. One in five reported a breach involving tools staff had adopted outside official channels. Organizations with high levels of shadow AI carried about $670,000 in additional breach cost.

Kyu Kwak, chief information security officer at Pearson plc, said the organizations that succeed will be the ones with “the strongest governance around autonomous systems.”

Employers are paying for the skills they say they cannot find. In Lightcast data cited by the consortium, U.S. cybersecurity jobs that require AI skills advertise a median salary 14.9% higher than the field overall.

Cisco set up the consortium in 2024, then called the AI-Enabled ICT Workforce Consortium, with nine other companies, among them Accenture plc, Google LLC and Microsoft Corp. Pearson joined this year, taking it to 11. The combined upskilling target is now 130 million people over the next decade.

A second spotlight report, on governing AI, is next.

Image: AI Workforce Consortium