The role of ITAD is changing as technology changes and evolves. Originally, it focused mainly on devices like laptops, desktops, storage, and servers. Still, with the rise of AI and edge computing hardware, these components are included in devices disposed of through ITAD. One challenge is that these assets are more expensive and harder to dispose of. ITAD providers are adapting to this new technology and adding new steps to their processes to dispose of AI and edge computing hardware properly. Let’s look more into why disposing of this hardware is different, what the risks are, and how ITAD providers are adapting to incorporate this technology into the disposal cycle.
What are AI/ML and Edge Computing Devices?
When we talk about AI/ML and edge computing devices, we refer to a variety of devices we continue to see as this technology develops. GPUs, TPUs, AI interface accelerators, and NPUs are all types of AI/ML devices that you might see. Edge devices include edge servers, IoT compute modules, and smart gateways, among others, found in healthcare, retail, and manufacturing.
Why These Devices Create a Challenge for ITAD
These devices differ and can often create challenges when it comes to asset disposition, especially when using current ITAD practices. Due to their high value, data security risks, and compliance pressures, there are multiple challenges associated with IT asset disposal for these devices. Let’s explore each of these challenges and why they exist:
Higher Value Assets
AI/ML and edge computing devices are considered “high-value assets,” and they can retain value differently than other devices that you might dispose of. On top of this, secondary market demand can be high but often changes rapidly, making it difficult to get into that market. Because these devices have a high value, there is also a push to reuse and recycle them for their materials, rather than just disposing of them and losing value.
Data Security Risk
With any ITAD device,e there is a potential data security risk, but with AI/ML and edge computing devices, this risk increases. More areascann store data, not just hard drives. For example, in these devices, data can be stored in embedded storage, system logs, cached files, and removable modules. Having all of these potential data storage areas makes it even riskier for data security and the potential for a data leak.
Compliance Pressure
With ITAD processes, there is a compliance standard that must be met, but when adding these new devices, that standard and the pressure become much more intense. The chain-of-custody requirements are growing rapidly, and they havecontinued to be an important part of ITAD, including for AI/ML and edge computing devices.
The Increased Data Risk and Why It Matters
We touched on the data risks associated with these devices above. Still, it is important to stress how much of an added data security risk these devicescann really pose. Edge computing devices store large amounts of data, and they can do so in various ways. For example, customer metadata, trained models, video footage, and even facial recognition data are all at risk in the event of a data breach. ITAD and data security threats look very different nowadays, thanks to the advanced technology we have. Artificial intelligence, in itself, stores huge amounts of data and personal information that can be detrimental if it falls into the wrong hands.
With the increased use of AI and the devices that come with it, there is also an increase in data security risks and the precautions needed to mitigate them. ITAD companies must understand these risks and put practices in place that can help avoid them.
How Traditional ITAD Processes Will Change
With the addition of these devices, ITAD processes are expected to change due to the specific needs for their disposal. The current standard process includes drive wiping, asset handling, and hand and server decommissioning, with the addition of disposing of the physical device. However, there are certain requirements for AI/edge devices that are specific to them and need to be implemented into modern-day ITAD processes. Here is a list of some of the additional processes that will need to be implemented into ITAD for these devices:
- Intake and sterilization of the hardware of these devices. For example, the VRAM size, model variant, and part numbers all need to be logged.
- Chain of custody should include limited-access handling, photo documentation of intake, and chain-of-custody logs at every handoff. These should be implemented because these devices have a high resale value, and they are more prone to theft.
- Functional testing for AI software will include stress testing, error reporting checks, and performance verification.
- Data sanitization must be expanded to include credentials, LGSs, datasets, and cached model files.
How to Prepare Your Organization
- Update our inventory to include AI and edge computing devices
- Build a specific checklist that’s just for AI and edge assets
- Make sure sanitization procedures are standardized across your organization
- Work with an ITAD provider that is familiar with AI and edge computing devices
- Create a policy that can be shared with your whole team for model/data handling during the disposal process
The Future of ITAD With AI Expansion
We can expect to see ITAD processes change significantly as artificial intelligence continues to grow and become more popular. This type of technology is expanding rapidly, making changes to ITAD processes crucial to keep up. We can expect tighter compliance requirements, greater demand for component tracking, and specialized testing for accelerators. As an organization, it’s vital to prepare for these ITAD processes to change and start implementing practices now. Reach out to an ITAD partner who is familiar with these devices to devise a plan that will help you stay on track and be prepared for the ever-changing technology landscape and the expansion of artificial intelligence.
