Industry Steps Up to Make AI Safe and Ethical

Given the emergence of AI-enabled products in the physical security industry, end-user customers, security integrators, MSPs, and consultants need to be aware of which companies are and are not addressing the critical global issue. 

Q:  IT is asking me to provide information about our use of AI in our physical security systems. What kind of information are they looking for?

A:  IT wants to know how using AI in physical security systems affects things like privacy for people on site, data security, security system cybersecurity risk, compliance with regulations, and whether there are safeguards in place to prevent AI errors from undermining security operations. They also want to know who is responsible for overseeing the use of AI. All this falls under what you could call “ethical use of AI considerations.”

Key Considerations

Many large companies now have IT review boards, including an AI Review Board, that qualify expansions and additions to physical security technologies, both on-premises and cloud-based. They will have questions about how AI in physical security systems impacts the following:

• Privacy rights of people on the site
• Data security, particularly regarding personal identifiable information (PII)
• Compliance with regulations related to these areas
• Whether safeguards exist to prevent AI errors from compromising technology performance
• Cybersecurity risk, including the attractiveness of AI-related computing capabilities in hardware devices (mainly cameras) and system software to hackers
• Whether safeguards exist to prevent AI errors from compromising security operations
• Who is responsible for overseeing and regularly testing how AI is being used

Among the concerns listed above, the first five bullet items generally align with areas that leading AI technology vendors for physical security—such as Axis Communications, Milestone Systems, and i-PRO—focus on as part of their ethical AI strategies. Their efforts emphasize privacy rights, transparency, fairness, accountability, and compliance with evolving regulations.

However, while these companies actively address many of these considerations, not all may fall squarely within their internal definitions of “ethical AI use.” For example, cybersecurity risks associated with AI-capable devices may be treated as a broader security engineering responsibility rather than as an ethical AI issue.

In practice, the last two bullet points—ensuring operational safeguards and defining oversight responsibility—fall directly to the organization deploying the technology. As such, these should be key discussion points between end-user customers, their security consultants, and their managed service providers to ensure appropriate governance, monitoring, and accountability for AI-enabled physical security systems.

Security Industry Ethical AI Use Leaders

Vendors such as Axis Communications, Milestone Systems, and i-PRO have taken visible leadership roles by establishing formal ethical AI principles, aligning with emerging regulations, and embedding privacy, fairness, and transparency into their AI-enabled offerings.

However, their definitions of “ethical AI use” are not all-encompassing and may not address every concern that an organization’s IT and risk management teams will raise—particularly issues like cybersecurity vulnerabilities in AI-capable devices.

Their leadership nonetheless sets essential benchmarks for the industry and signals that ethical AI is becoming a differentiator among physical security vendors. For end-user organizations, these developments underscore the importance of asking informed questions, looking beyond vendor claims, and taking ownership of how AI-enabled systems are governed and overseen within their environments.

Axis Communications

Axis Communications has made responsible AI a central part of its corporate strategy. The company emphasizes respect for human rights, privacy, and transparency as foundational principles, insisting that responsible AI must be embedded in product development from the outset rather than being treated as an afterthought. Axis actively monitors regulatory developments, such as the EU AI Act, to ensure compliance and promote industry best practices.

“Our approach to AI is rooted in one overarching principle: That AI technology, just like all technologies, should leverage and augment human intelligence, build on respect for human rights and should benefit people and society.” — Mats Thulin, Director of AI & Analytics Solutions, Axis Communications, echoing a company principle explained in the Axis AI in video surveillance (2025) report.

i-PRO

Formerly Panasonic Security, i-PRO has formalized its ethical AI approach through its own Ethical Principles for AI, which aim to strike a balance between innovation and societal responsibility. The company has achieved ISO/IEC 42001 certification for its AI management systems and has committed to transparency, privacy protection, and continuous improvement of its ethical AI practices. I-PRO positions itself as advancing safety and security while respecting the rights and privacy of individuals.

Milestone Systems

Milestone Systems was the first video management software company to adopt the G7 Code of Conduct on Artificial Intelligence, underscoring its commitment to trustworthy and responsible AI globally. The company integrates ethical considerations throughout the AI development lifecycle, encompassing research on societal impact, internal testing, post-deployment monitoring, and prioritizing transparency and human oversight. Milestone also leads collaborative efforts to shape responsible AI standards in the industry.

“We need rules to ensure AI is being developed to serve humanity. But companies should not wait for regulations. They must take their steps to identify and resolve the weaknesses and pitfalls of the AI they develop.” — Thomas Jensen, CEO, Milestone Systems.

Security Industry Association (SIA) Leadership

After nearly two decades of criticizing the Security Industry Association’s convergence efforts regarding what I called “shallow IT adoption,” I am pleased—and probably overdue—to report that the SIA has been increasingly active in addressing emerging technology issues, particularly those related to AI, privacy, cybersecurity, and governance. Today, SIA is not only keeping pace with technological advancements comprehensively but is also beginning to play a significant leadership role, thanks in part to the active participation of experienced individuals from companies such as Axis Communications, Milestone Systems, and i-PRO.

SIA’s current efforts reflect a broad approach to the ethical and responsible use of AI in physical security. The association now advocates for privacy-respecting frameworks, strong PII protections, regulatory alignment with frameworks like the EU AI Act and the NIST Risk Management Framework (RMF), heightened cybersecurity standards for AI-capable devices, continuous monitoring to ensure AI performance reliability, and clear governance structures for oversight and accountability.