As the telecommunications landscape gears up for the advent of 6G technology, the role of artificial intelligence (AI) becomes increasingly indispensable, bringing with it immense opportunity for innovation as well as major challenges for the wireless industry. On one hand, AI promises to usher in an era of improved mobile network efficiency. However, testing AI-integrated devices requires new testing methodologies that push the boundaries of measurement science. On the other hand, as devices continue to increase in complexity, meeting time-to-market, quality and cost objectives will become increasingly complex. In this space, AI can help acquire more actionable insights from measurement data to help make business decisions. This strategic embrace of AI not only ensures seamless integration into the dynamic wireless landscape but also lays the foundation for future breakthroughs, propelling the industry toward a future defined by innovation and excellence.


The convergence of AI and testing unfolds through a multifaceted lens, encompassing diverse applications and transformative potentials. However, there are two main use cases driving change in the industry. First, generative AI begins to emerge as a tool for accelerating design workflows, optimizing time-to-market and minimizing operating costs. Second, wireless industry players are increasingly integrating AI into their products, leading to an increased device under test (DUT) complexity requiring new testing methodologies.


Figure 1

Figure 1 Product development life cycle.

As of 2024, wireless connectivity has become ubiquitous, reaching most households and mobile devices worldwide. This widespread adoption signals a shift in industry dynamics from a primary focus on revenue growth to one centered on profitability and optimization. This transition towards optimization has propelled discussions surrounding the integration of AI in 6G wireless technology as industries seek to enhance efficiency and maximize returns in this mature market landscape. In the context of test and measurement, AI-driven solutions offer unparalleled opportunities for streamlining workflows, optimizing resource allocation and enhancing overall operational efficiency. Figure 1 shows a product life cycle flow, including some of the compounding and conflicting issues that pressure this product cycle.

The integration of generative AI into design workflows marks a transformative leap for the industry. By leveraging AI-driven insights derived from measurement data, businesses can unlock unprecedented levels of productivity and competitiveness. Generative AI algorithms play a crucial role in optimizing the design of intricate antenna systems, enabling rapid prototyping and iterative refinement to meet stringent performance requirements. The need to realize these benefits is becoming much more concrete, as evidenced by the data points shown in Figure 2. These two results come from internal NI research conducted in partnership with the research division of the Financial Times Group.1

Figure 2

Figure 2 Findings from an NI 2022 research report.

This strategic convergence not only expedites design workflows but also empowers engineers and designers to innovate. In sectors like the semiconductor industry, AI-driven design tools revolutionize chip layout optimization, yielding higher performance and energy efficiency while shortening design cycles. Ultimately, the integration of AI into design processes presents unparalleled opportunities for enhancing productivity, efficiency and innovation, propelling industries toward sustained growth and competitiveness in the dynamic landscape of wireless technology.


Figure 3 shows a simplified rendition of the workflow involved in characterizing and validating a device. The process typically commences with defining the desired outcomes, which may stem from various sources such as design specifications, product requirements or cost and time constraints. Engineers then translate these requirements into a comprehensive test plan encompassing all necessary tests to evaluate the device against the specified criteria. Subsequently, they develop and refine the tests, striving for maximum automation. This phase often consumes the most amount of test time.

Figure 3

Figure 3 Simplified device characterization and validation workflow.

Following test development, engineers execute the tests on multiple devices to assess individual performance and device-to-device variations. The results are then meticulously analyzed and reported with room for refinement or optimization of tests and sequences as needed. Depending on the complexity of the device, this process may extend from weeks to months, exerting pressure to expedite it to accelerate revenue generation.

By leveraging state-of-the-art software and hardware tools, like NI’s LabVIEW and TestStand, organizations can significantly boost engineering efficiency within each step of the process. NI has developed a prototype of a significantly optimized workflow driven by generative AI. This approach utilizes an AI infrastructure with a chat interface to autonomously generate tests and sequences based on given requirements and datasheets. Automating this segment alone could slash the time required for design characterization from weeks to days. The integration of tools like generative AI marks another significant leap forward, promising to expedite time-to-market, reduce operating costs, enhance leverage and promote reusability.


The integration of AI into 6G wireless products ushers in a new era of complexity and innovation. Unlike their traditional counterparts, AI-enhanced DUTs present a host of unique challenges that demand innovative solutions to guarantee both performance benefits and trustworthiness. For instance, in telecommunications, AI-driven network optimization algorithms are deployed to enhance spectral efficiency and minimize interference, thereby elevating overall network performance and user experience. This widespread adoption of AI underscores the pressing need for the test and measurement industry to evolve and craft specialized solutions tailored to the testing requirements of AI-enhanced DUTs, ensuring seamless integration and optimal performance in real-world scenarios.

As the industry gears up for the era of 6G and beyond, complexity reaches unprecedented levels as devices integrate advanced functionalities and AI-driven intelligence. The dominance of software components over hardware further amplifies the demand for frequent testing to maintain reliability amidst rapid software evolution. The integration of AI exacerbates this complexity, necessitating rigorous evaluation to guarantee safety and trustworthiness. Given the pivotal role of these devices across industries, ensuring quality remains paramount, prompting new test challenges like efficiently sourcing (the right) data, setting up accurate scenarios for test and managing the “infinite test space.” Some details of these challenges are shown in Figure 4.

Figure 4

Figure 4 Embedded, trustworthy AI in 6G.

While the technology of AI applied to wireless is new and with many challenges, addressing them allows the wireless industry to effectively harness the transformative power of embedded AI. In applications where spectrum, energy and chip real estate are both finite and valuable resources, AI can help optimize them in next-generation wireless devices. This helps pave the way for enhanced performance, reliability and efficiency in future wireless networks.


The integration of AI into test and measurement processes stands at the forefront of a transformative era, poised to redefine industry standards and methodologies. This is particularly true in the realm of 6G wireless and beyond. This strategic embrace of AI not only ensures seamless integration into the dynamic wireless landscape but also lays the foundation for future breakthroughs, propelling the industry toward a future defined by innovation and excellence.

AI algorithms are poised to play a pivotal role in revolutionizing test management processes, automating scenario selection, optimizing test coverage and mitigating the complexities associated with diverse use cases and edge conditions. However, testing AI devices in the context of 6G applications presents formidable challenges, particularly in the critical phase of AI model training. This phase, essential for creating robust and reliable AI models, encompasses three main steps: model design, training and validation. While synthetic training data generated through simulation tools offer some utility, the accuracy of this training data hinges on the fidelity of the simulation models. In contrast, real-world training datasets, acquired under authentic channel conditions, offer superior quality but are inherently more challenging to obtain due to the specialized hardware and software required for data recording. Striking a balance between these considerations is essential to ensuring the robustness and reliability of AI models tailored for 6G applications.


Traditional stimulus-response testing systems fall short when it comes to validating embedded AI devices. The challenge lies in the unpredictable nature of machine learning-trained systems, which may exhibit unexpected behavior across a broad spectrum of test scenarios. Unlike traditional algorithms, embedded AI models can be highly sensitive to environmental factors and system configurations, making it challenging to identify relevant stimulus signals. To address this, scenario-based test systems offer a viable alternative, abstracting different test conditions into easily understandable scenarios with predefined parameters and outcomes. These scenarios are then dissected into concrete test cases for comprehensive evaluation. Despite the potential complexity of scenario selection, leveraging smart techniques enables the identification of pertinent scenarios, ensuring thorough testing of AI devices in the wireless space. As the wireless landscape continues to evolve, scenario-based testing is a promising approach to efficiently validate the performance and reliability of embedded AI technologies.


The “infinite test space” refers to the vast array of potential scenarios and conditions that an AI system may encounter in real-world applications. Unlike traditional software testing, where inputs and outputs can be exhaustively enumerated, AI systems are trained on data and may exhibit unexpected behaviors when faced with novel situations. This means that testing AI involves grappling with an expansive and often unpredictable range of circumstances, making it impractical to test every possible scenario. Hierarchical scenario descriptions, coupled with smart scenario reduction techniques, are needed to manage the complexity of the test and measurement process while maintaining the test coverage and ensuring the reliability and robustness of an AI-enhanced DUT.


The integration of AI into wireless applications represents a pivotal advancement, with discussions around its role in future standards gaining momentum. Notably, the 3rd Generation Partnership Project (3GPP) is actively exploring AI’s integration into the forthcoming 5G Advanced and 6G standards, a topic of significant interest in Release 18 and 19 discussions. This strategic embrace of AI is not merely speculative but grounded in pragmatic considerations, driven by the imperative to enhance profitability within the wireless industry. Given the scarcity of resources such as spectrum, size constraints and power consumption concerns, even marginal improvements facilitated by AI can yield substantial cost savings. AI holds the potential to unlock significant gains across various fronts, including spectral efficiency, interference reduction, chip size optimization and power consumption, thereby reshaping the landscape of wireless technology. Figure 5 shows a base station tower and sectors, along with wireless backhaul links, as an example of a network that can benefit from AI integration.

Figure 5

Figure 5 Wireless base station tower, sectors and backhaul installation.

While AI’s current application primarily targets higher network layers, its potential extends to lower layers of the protocol stack, presenting a burgeoning trend in 6G discourse. Research endeavors are actively exploring the integration of AI into lower layers, recognizing its capacity to improve spectrum and energy efficiency as well as performance. However, to realize the promise of embedded AI within wireless networks, several prerequisites must be met. These include access to a sufficient quality and quantity of training data, the development of test systems capable of emulating real-world scenarios and the implementation of robust methodologies for navigating the infinite array of testing scenarios.


The integration of AI into the physical layer of 6G devices represents a monumental leap in wireless communication technology, promising to unlock new possibilities and transform industries. However, with great innovation come significant new challenges, particularly in the realm of testing. The industry must proactively address the test implications of AI integration to ensure that 6G devices deliver on their promises of ultra-low latency, reliability, energy and spectrum efficiency, massive connectivity and blistering data rates.

As we move closer to the era of 6G, collaboration, standardization and the development of AI-powered testing solutions will be critical. Establishing industry-wide standards for testing 6G devices with embedded AI is paramount. These standards should encompass AI algorithms, AI training and corresponding data, as well as performance metrics and testing methodologies, to ensure consistency and reliability. Collaboration among device manufacturers, network operators and AI experts is essential to address the challenges of AI integration. Sharing best practices and insights can lead to more robust testing methodologies, as well as ensuring interoperability with optimum performance. By addressing these challenges head-on, we can help ensure that AI-integrated 6G devices operate flawlessly, ushering in a new era of connectivity.