The latest research from PSA Certified underscores an increasingly critical emphasis on security as artificial intelligence (AI) continues to expand. This study reveals that a substantial 68% of technology decision-makers express apprehension that the rapid advancement of AI could surpass the industry’s current capabilities to secure products, devices, and services. This highlights the need for parallel advancements in security investment and best practices to ensure that AI deployment is trustworthy.
One significant factor enhancing the urgency for better AI security is the proliferation of edge technology. With edge devices capable of processing, analyzing, and storing data locally—right at the device or network edge—they offer efficiency as well as security and privacy benefits over centralized, cloud-based approaches. Therefore, it comes as no surprise that 85% of device manufacturers, design manufacturers, SIPs, software vendors, and other technological decision-makers predict that security concerns will drive more AI applications to the edge.
However, with the race for efficiency in full swing, the security of edge devices has become more integral. Today’s organizations face the critical task of bolstering the protection of their devices and AI models to meet the challenges of deploying AI at scale. According to the report, the security considerations span the entire supply chain. A survey involving 1,260 global technology decision-makers reveals that over the past year, security has grown as a priority for 73% of respondents, with 69% attaching increased significance to security as AI adoption grows.
Despite AI’s potential to elevate security, the report uncovered a lag that must be addressed to harness AI’s full potential. Surprisingly, only 50% of the survey participants believe their current security investments are adequate. Many are neglecting foundational security measures such as security certification, which is pivotal for establishing best practices.
Currently, just over half of the respondents are utilizing externally validated security certifications (54%), independent third-party testing or evaluation (48%), or threat analysis and threat modeling (51%) to enhance the robustness of their products and services. These fundamental security strategies are vital as more businesses endeavor to increase consumer trust in AI-driven services.
David Maidment, Senior Director of Market Strategy at Arm—a co-founder of PSA Certified—emphasized the pivotal relationship between AI and security. “One doesn’t scale without the other,” he stated. “AI presents immense opportunities, but it also offers those opportunities to malicious actors. It’s crucial now more than ever that those within the connected device ecosystem adhere to best practice security and not compromise in their quest for AI functionalities. The entire value chain must collectively assume responsibility to maintain consumer trust in AI services. The positive news is the industry’s recognition of the necessity to prioritize security investments to future-proof systems against new attack methods and rising security threats linked to rapid edge AI adoption.”
Interestingly, four in five respondents recognize that incorporating security into products not only drives consumer confidence but also benefits the bottom line. The importance of security regulation compliance has similarly risen, underscored as a top priority by 80% of respondents, an increase from 74% the previous year.
With edge AI surging in conjunction with a spike in AI inference, a substantial volume of personal data is now processed across billions of endpoints, each necessitating secure protocols. To handle the security of edge devices and ensure compliance with emerging cybersecurity regulations, stakeholders within the connected device ecosystem must collectively partake in establishing a secure edge AI lifecycle. This encompasses not just secure device deployment but also the secure management of trusted AI models at the edge.
Despite concerns that AI advances might eclipse industry capabilities to secure products, the sentiment remains optimistic. A significant 67% believe their organizations are well-prepared to manage the security risks posed by escalating AI integration. Interestingly, more decision-makers prioritize enhancing the security of their products and services (46%) over increasing their AI readiness (39%), recognizing the necessity of scaling both security and AI concurrently.
Nevertheless, with a notable 78% of respondents acknowledging the need for further AI preparation, security must stay at the forefront of technology strategy. Scaling security in an era of interoperability and edge AI demands established standards, reliable certification, and trusted hardware. Embedding security by design will afford organizations a best-practice benchmark, shielding them against future risks.
“To unlock AI’s full potential, stakeholders must mitigate potential security risks,” concluded Maidment. “As the connected device ecosystem embraces AI-enabled use cases, it is crucial to progress without disregarding security implications. Security must be an intrinsic blueprint, embedded across the entire value chain to maintain best practice and pace with evolving security threats.”