For the public sector—including governments at the federal, state, and local levels—investing in AI offers tremendous benefits. AI has the potential to help organizations work efficiently, manage costs, and make great progress in research.
Private companies have been making incredible advances in AI. By applying what they’ve learned to challenges in the public sector, everyone benefits from what AI offers. Similarly, government is often a driver of some of the most important technology advancements. Government and public sector innovation can help promote the adoption of AI across industries.
Benefits of AI in Government
At the heart of AI in government services are techniques like machine learning and deep learning, computer vision, speech recognition, and robotics. When put to work, these techniques translate into real, tangible benefits. Natural language processing automatically extracts relevant information from intelligence sources and makes connections, enabling analysts to find actionable insights. Failure rate predictions help ensure military equipment is well maintained and ready to deploy when needed. Cyber anomaly detection can revolutionize cybersecurity strategies in government systems. The possibilities are endless and yet only beginning to take shape.
Automate Routine Tasks
The human workforce cannot manually process large amounts of data, particularly the massive sets collected by the military, aerospace, and other government sectors. But AI completes time-consuming—and tedious—tasks quickly and accurately. Machine learning and deep learning in particular excel at tasks that are clearly defined, involve massive amounts of data, and require extensive repetition.
By automating simple, well-defined tasks, AI streamlines operations and augments the workforce. Employees can then spend more time on decisions that require human input. For example, the U.S. Citizenship and Immigration Services (USCIS) uses a computer-generated virtual assistant named Emma to answer questions and direct individuals to the right area of the website. But a human can better explain USCIS decisions and coach an individual on next steps.
Perform Mission-Critical Tasks
In government agencies like the U.S. Department of Defense (DoD), AI can support critical capabilities. These include improving situational awareness and decision-making; increasing the safety of equipment like aircraft, ships, and vehicles in dangerous situations; and predicting when critical parts will fail, automating diagnosis, and planning maintenance. AI also helps improve the nautical, terrain, and aeronautical charting vital to DoD missions, helping enable safe and precise navigation and better surveillance.
Advance Research and Development
The combination of AI, high performance computing (HPC), and analytics powers research in medicine, genomics, engineering, seismology, astrophysics, and many other areas. Government-funded research is especially poised to benefit from AI techniques in academia and the private sector. For example, AI algorithms enable faster and cheaper predictions of 3D protein structures from sequences of amino acids.1 This knowledge may help researchers better understand diseases and design drugs to treat them.
Intel takes a holistic, systems-level approach to AI. This view helps ensure organizations deploy the right system with the performance needed now and going forward.
Intel® AI: A Systems-Level Approach
For AI to work, it’s critical to have the right data, ensure it’s accurate, and label it properly for training. Models must be trained in the right software and hardware environment. This is especially true when deploying AI models in the field. For example, AI algorithms running on a satellite in orbit should learn from new data and continually adapt their models. It’s also important to consider the AI system as a whole, including compute, memory, storage, and interconnect, to prevent any one part from causing a bottleneck.
Intel takes a holistic, systems-level approach to AI. We work with government and public sector teams to determine how to manage and prepare data, mature and refine models, and deploy hardware and software that meet field requirements. This view helps ensure organizations deploy the right system with the performance needed now and going forward.
To meet public sector budgets, scalability and TCO are also crucial. Intel® Xeon® Scalable processors support not only AI workloads, but also a broad range of general-purpose workloads. This maximizes the utilization of infrastructure and can lead to a greater willingness to invest in new hardware.
The full portfolio of Intel® technology for AI offers a range of additional tools, development kits, and solutions for use with cloud service providers.
Intel® Technologies for AI from Edge to Cloud
The Intel® portfolio of technologies for AI supports multipurpose, purpose-built, and application-specific systems. This means AI teams can select the best hardware for each job and optimize with software to get the maximum performance. The following components provide a strong foundation for almost any AI application.
High-Performance Hardware |
|
|---|---|
CPUs |
Intel® Xeon® Scalable processors deliver high performance for machine learning and deep learning, plus high utilization to provide more value for infrastructure investments. |
Accelerators |
Intel’s portfolio of accelerators are designed to handle specific, rigorous AI workloads. These include Intel® FPGAs for the data center, network, or edge and Intel® Movidius™ Vision Processing Unit (VPU) for computer vision. |
Memory and storage |
Intel® Optane™ persistent memory and SSDs overcome bottlenecks in data movement and support large in-memory applications. |
Interconnect |
Intel® technologies for interconnect scale from microns to miles, spanning on-die, on-package technology to data center and wireless networks. |
Software and Developer Tools |
|
|---|---|
Framework optimizations |
Optimizations for popular frameworks like TensorFlow and Caffe can help developers boost performance on Intel® architecture. |
Software libraries |
Software libraries, including Intel® Math Kernel Library for Deep Neural Networks (Intel® MKL-DNN) and Intel® Data Analytics Acceleration Library (Intel® DAAL), help streamline development. (See optimization notice.) |
Software toolkits |
Software tools like the Intel® Distribution of OpenVINO™ toolkit let AI developers create a model and deploy it across different types of Intel® hardware at the edge and in the data center. |
Government organizations that build their AI systems on a foundation of Intel® technologies gain the advantage of hardware-enabled security features.
Hardware-Enabled Security Features
Not all technology providers develop solutions with security in mind. Government organizations that build their AI systems on a foundation of Intel® technologies gain the advantage of hardware-enabled security features. These help secure AI algorithms and data streams from the endpoint—whether that’s a ruggedized laptop, drone, or satellite—through the network and to the cloud.
Our technologies that enhance hardware-enabled security include Intel® Threat Detection Technology (Intel® TDT) to improve the awareness of cyber threats and exploits. In addition, Intel® Security Essentials includes Intel® Trusted Execution Technology (Intel® TXT) to help ensure applications run in secure enclaves; hardware-assisted acceleration of crypto operations; protected data, keys, and identity; and a protected and verified boot process.
Case Studies: Intel® AI in the Public Sector
From data strategy to enterprise-scale deployment, Intel and our ecosystem are helping speed time from pilot to impact with production-ready solutions for both the private and public sector. The following AI in government examples show how these solutions can bring enormous value to the public sector.
More than 700,000 residents depend on the District of Columbia Water and Sewer Authority for reliable wastewater removal. But with parts of the infrastructure more than 200 years old, the city’s 1,800 miles of sewer pipe need continual inspection and maintenance. Video analysis with Intel® AI technologies helps cut costs and optimize maintenance.
Every year, the FBI reports almost half a million missing children. The nonprofit National Center for Missing & Exploited Children serves as the clearinghouse for critical information to help find the children. Advanced analytics and AI can help process tips about suspicious online activity and quickly get the information to the appropriate law enforcement agency.
Defense Advanced Research Projects Agency (DARPA)
To drive advancements in AI research, Intel participates in DARPA challenges and DARPA artificial intelligence programs. For example, Intel and Brown University are collaborating on Intelligent Spine Technology. The program explores the use of Intel® AI technology, including accelerators and open source software, to treat spinal cord injuries and help paralyzed patients regain movement. Researchers plan to record motor and sensory signals and learn how to use artificial neural networks to stimulate the postinjury site.
The Future of AI and Government
Intel is exploring new techniques and strategies for government digital transformation. For example, federated learning aided by Intel® Software Guard Extensions (Intel® SGX) may allow multiple institutions to collaborate on projects without sharing private data.
For every organization, AI is a journey. But AI offers myriad benefits for governments and the public sector to accomplish more, satisfy employees and citizens, and achieve unimagined advances. Organizations can prepare for success by carefully planning each stage with the help of experts in technology and deployment. We’re still at the beginning with AI in the public sector, and it’s time to prepare. The future has arrived.