Artificial intelligence (AI) adoption is expanding across manufacturing organisations, but increased use of AI applications and connected systems is creating new cybersecurity and data protection challenges, according to a report by Netskope Threat Labs.
The Netskope Threat Labs Report: Manufacturing 2026, released on November 10, found that the proportion of manufacturing users actively using AI applications increased from 40 per cent to 67 per cent over the past year, reflecting the technology’s growing integration into business operations.
The report also identified a shift towards organisation-managed AI tools, with their adoption increasing from 35 per cent to 70 per cent. In comparison, the use of personal AI accounts declined from 75 per cent to 55 per cent over the same period.
Microsoft Copilot was the most widely adopted AI application among manufacturing organisations, used by 90 per cent of organisations, followed by Claude Platform at 89 per cent and ChatGPT at 76 per cent.
However, the findings highlight growing concerns about how organisations manage information shared with AI systems. Regulated data accounted for 57 per cent of AI-related data policy violations, followed by passwords and access keys at 19 per cent, intellectual property at 16 per cent and source code at nine per cent.
AI integration expands beyond standalone applications
The report found that AI adoption extends beyond employees directly accessing dedicated AI tools. Some 96 per cent of manufacturing employees use applications with embedded AI features, while 90 per cent interact with systems that specifically use customer or user data for model training.
Netskope Threat Labs said this broader integration makes it more difficult for organisations to track how sensitive information is shared and processed across their technology environments.
The use of agentic AI systems is also increasing. The report recorded a 490 per cent rise in the number of non-human entities connecting to remote Model Context Protocol (MCP) servers, alongside a 600 per cent increase in MCP-related events.
MCP is an open-source standard that allows AI models to connect with external and internal data sources and tools. According to the report, these connections can provide AI agents with access to business information and other sensitive resources, creating additional security considerations for organisations.
Malicious AI activity adds to security concerns
The research also identified increasing malicious activity targeting employees through AI-related content and applications.
The number of manufacturing users encountering malicious AI lures reached nearly 175 per 100,000 employees in June 2026. These lures typically involve fake AI applications, installers or developer tools designed to encourage users to download malware or visit malicious websites.
Interactions with malicious links originating from AI applications also increased, reaching 200 per 100,000 users in March 2026.
Netskope Threat Labs said organisations should apply security checks to links returned by AI applications, similar to those used for links found through search engines and email.
The report recommends that organisations strengthen monitoring of web and cloud traffic, restrict access to applications that present disproportionate security risks, and implement data loss prevention policies to identify and protect sensitive information.
It also recommends improving visibility and control over AI applications, including those operating through personal accounts and enterprise environments.
The findings are based on anonymised, aggregated usage data from a subset of Netskope customers in the manufacturing sector, covering the period from June 1, 2025, to August 31, 2026.












