Realizing the Full Potential of Edge AI with Connected Security

AI has taken the world by storm. The number of connected devices with AI-based computing capabilities is booming. However, this massive growth raises security concerns for businesses that cannot be ignored. As AI increasingly makes automated decisions with no human oversight, ensuring the security and integrity of AI systems within the billions of connected edge devices is vital for businesses and consumers to safeguard their personal data and trust the AI-based services that they are using.  
 
To understand security challenges in the AI era, the PSA Certified 2024 report, produced in partnership with Arm, examines survey results from 1,260 global technology decision-makers. The findings reveal that over the past 12 months, security has become a higher priority for 73% of respondents, with 69% specifically emphasizing security due to AI advancements. In addition, 71% agree that AI-driven cyber threats are a significant cause for concern. Yet, over a third (33%) of technology decision makers are unaware if their organization is fit for purpose in the age of AI. Overall, key survey findings indicate a gap in AI-security integration that must be addressed so organizations can fully realize AI’s potential.  

Maintaining AI innovation while keeping pace with security

Despite security concerns, 67% of respondents agreed that the benefits of AI still outweigh the risks. While organizations are eager to capitalize on the AI opportunity and are excited about its potential, they don’t want to put their organizations at risk. Whether it’s obtaining the proper product and component certification or doing proper security threat modeling, roughly only half of businesses have the basics covered when it comes to improving the security of their products and services.  

Democratizing best practice security to ensure a baseline across the ecosystem will be critical to the scaling of AI. Securing devices, ensuring regulatory compliance at the edge and enabling secure AI model development are all challenges facing stakeholders in the connected device ecosystem. This is especially the case for regulatory compliance, as governments are becoming increasingly aware of the reliance that economies have on digital services, with additional moves to regulate security best practices in schemes like the EU Cyber Resilience Act and UK Product Security and Telecommunications Infrastructure regime. The rapid proliferation of AI cannot happen in isolation; it needs to be matched with trust, which can only happen by investing in security best practices.

Focusing on security in the age of AI 

With billions of diverse AI-enabled connected devices processing unprecedented amounts of personal data, every customer expects their data to be secure. For businesses, the challenge is how to protect and verify AI data from the cloud to the device. This requires security at scale and an evolution of edge computing. In fact, 85% of respondents believe that security concerns will drive more AI use cases to the edge, highlighting the magnitude of the task ahead. 

However, the ability to process, analyze and store data at the edge of the network has significant advantages over a centralized cloud-based location. These edge AI benefits are recognized by the survey respondents, with 51% highlighting the improved efficiency, 50% pointing to better security and 41% stating faster data processing and reduced latency. 
 
A common issue in security preparedness is that organizations are not adopting a security-by-design approach as a standard practice. As a result, they miss the benefits of embedding security into a device from the outset. This problem is magnified by the fragmented nature of the supply chain. Multiple companies are involved in device creation, and without a unified resource or framework, this fragmentation persists, leading some organizations to overlook security by design, putting the entire ecosystem at risk.  

The PSA Certified security ecosystem

It’s crucial to implement security measures from the ground up, with all OEMs prioritizing certification. In fact, 80% of technology decision makers say that building security into their products is a driver of improved market acceptance and trust. Moreover, just over half (54%) are currently using externally validated certifications, independent third-party product testing (48%) or threat analysis / threat modeling (51%) to improve the security robustness of their products and services.  
 
PSA Certified is actively tackling this issue. Founded by Arm and six other security ecosystem leaders, PSA Certified is a global partnership of security-focused companies dedicated to integrating security best practices into devices at scale. By independently evaluating and certifying the components and chips of connected devices, companies can ensure security best practices are upheld throughout the supply chain. Certified chips and certifications can be reused, allowing expertise to be shared broadly across the value chain.  

Prioritizing security at the edge

Another major factor contributing to greater AI-security is edge technology and investing in hardware to future-proof systems against new attack methods and satisfy security imperatives. With the ability to process, analyze and store data at the edge of the network, or on the device itself, deploying AI at the edge can enable improved privacy, lower latency and greater reliability. However, there is still a need to ensure that edge-based devices are secure and maintain secure connections to cloud-based services.  
 
To meet the demands of deploying AI at scale, organizations must intensify their data protection efforts, making the security of edge devices more critical than ever. The current security landscape is viewed positively by survey respondents., 70% believe that the electronics industry is developing trusted hardware suitable for the AI era, and 71% trust their chosen hardware to protect their intellectual property, including valuable AI models that are deployed on edge devices.    

Ensuring trusted AI deployment

Though decision-makers across the connected device ecosystem agree that the benefits of AI will outweigh the security risks, the rapid growth and innovation in AI connected devices needs to be matched by investments in security best practices to ensure trusted AI deployment. Identifying and prioritizing a security strategy and investment in connected devices is essential for protecting both businesses and their customers.

Read the full report

You can gain more insights into connected device security in the age of AI by reading the PSA Certified 2024 Report: Scaling Security in the Age of AI.