Imagine a United States where Cybersecurity and Infrastructure Security Agency (CISA) can predict and neutralize global ransomware attacksi, Customs and Border Protection (CBP) can anticipate and intercept methamphetamine dropsii, and National Institutes of Health (NIH) can prognosticate longitudinal disease response based on clinical trial results. Achieving these efforts will result in a nation that not only counters emerging threats but also is ready to leverage AI for the benefit of its citizens’ health, safety, and overall well-being. In this ever-evolving digitally transformed landscape, our citizens’ data would be better protected, criminal activity would be more effectively controlled, and scientists and patients can achieve better health outcomes at lowered costs. All these technological improvements are now possible with Artificial Intelligence (AI).
AI is a rapidly evolving field with the potential to transform businesses of any size and industry – be it private or public. The 2022 surveys from the Center for Digital Government’s Digital Counties, Digital States, and Digital Citiesiii demonstrate AI’s strong position among local and state government as chatbots – majority of agencies are either already using them or plan to use them in the next 12-18 months.
Artificial Intelligence – A Product of Human Creativity
Artificial Intelligence covers complex systems designed to mimic human behaviour and intelligence. It achieves said intelligence by processing massive amounts of data quickly and making connections between them to glean meaningful insights and perform actions that may otherwise need human intervention. Intelligence is often characterized by our learning speed, accuracy of our understanding, and the logical soundness behind connecting the information.
Building an AI is the science of training machines to do complex tasks that are considered “smart,” like problem-solving and decision-making. From simple tools that responds to customers via chat to complex machine learning algorithms for data analysis, AI comes in many forms and has a diverse range of purposes. Fundamentally, AI systems combine large data using sophisticated algorithms to analyse, understand, and make decisions or predictions about future states. While its capabilities are still limited to programming, a part of its sophisticated algorithm during programming is the adaptability to consume a growing data set, incorporate it as a part of its processing, identify patterns from the aggregate dataset, make inferences based on those patterns, and then, take action. This allows the machine to improve its performance over time, just like humans learn from experiences.
However, much of its potential is still dependent on not only how flexibly its algorithm was set – which requires programmers to write prescient code that allows for acceptable responses during unanticipated conditions – but also the training dataset that becomes the foundation for its knowledge base and ability to generalize. This foundational component becomes a critical determinant so the thinking and planning around the diversity and quality of the seeding dataset and how to program it for adaptivity and scalability is dependent on the level of human creativity behind its build.
Training Your AI in Depth
The concept of AI is not new; it dates back to mythology across cultures as humanoid creatures constructed out of inorganic materials have played roles as guardians, anti-heroes, obstacles, and helpers. The breakthroughs that made the current form of AI that mechanizes human thought possible are recent progress in computational power and deep learning models. Deep learning is a type of machine learning framework that uses artificial neural networks to process data and scales the machine’s learning algorithm with the data volume by using a layered architecture and labelled data. Inspired by the structure and function of the human brain, artificial neural networks can learn complex patterns in data. Deep learning enables AI systems with the capability to automatically learn and extract meaningful information from data, making it an essential element of modern AI applications. Its ability to extract and organize complex features and strength in scaling with data volume have made the following contributions to AI advancements:
Image Recognition: Image recognition systems are using deep learning models to identify objects and scenes in images. These systems and used in applications such as self-driving cars, facial recognition systems, and medical imaging systems.
Natural Language Processing (NLP): The model is used in NLP systems to understand and generate human language and are used in applications such as – machine translation, chatbots, and voice assistants.
Machine Translation: Machine translation systems use the model to translate text from one language to another. Businesses and individuals use the model to communicate with people who speak different languages.
Feature Extraction: Deep learning algorithms automatically identify and extract relevant features from raw data, allowing AI systems to understand and process information effectively.
These models hold powerful extraction and classification capabilities that deliver advanced results, enabling cutting-edge processing to include image recognition, natural language, and machine translation.
Impact of AI
According to Forbesiv, the amount of data created, captured, copies, and consumed increased by almost 5000% between 2010 to 2020. So, what are we going to do with so much data? With AI, companies can now capture these data to increase their business goals and value by enhancing end user experiences. AI is driving so much change which makes it a necessity to embrace the technology and automation and make it easier for people to get things done.
Our Technology Partners at Work
NetImpact is an innovation leader with a proven track record in driving modernization success across Federal civilian and defense agencies. However, we don’t do this alone. Our DX360°® appsv are Federally bespoke COTS built natively on industry-leading platforms – extending the best-in-class investment our clients have made. While the powerful DX360°® apps come unboxed compliant with best practices, Federal policy, and other guidelines, they leverage the underlying ServiceNow and Microsoft platforms. Our Partner’s AI investments become an inherited benefit for our products – below are just two small ways they’re pressing forward:
ServiceNow is doing a lot of work with AI in the private as well as in the public sector with several Federal agencies. Their intelligent platforms work continuously towards accelerating digital transformation by bringing all the capabilities in one spot making it easier for everyone to work towards a common goal with a focus on three vectors: experiences, speed, and business transformation.
ServiceNow is really thinking about how its service impacts the end users. Chris Bedi, Chief Digital Information Officer, ServiceNow recently unveiled a service desk scenario that would improve the lives of airmen: an airman initiates a Virtual Agent conversation after realizing the warehouse she is working is becoming unbearably warm and distracting. Here’s what happened:
The airman reports, using natural language, that “It’s very hot in here!” The Virtual Agent walks through some standard protocol questions to troubleshoot, such as requesting confirmation that the windows are closed. It gets interesting from there: the Virtual Agent can then identify the four Air Conditioning (AC) units located, locate the team closest in proximity, and leverage API calls to check, then reset the thermostat. When the room temperature fails to drop from 81° in the sweltering, southern August heat – the Virtual Agent prioritizes a technician appointment. Pretty “cool.”
Microsoft has been working on the new era of AI for the last couple of years. Dating back to 1956, our partner began by figuring out how computer science can help make machines smart enough to be perceived as having intelligence. Fast-forward to the 90s and current times, their technology reached the level of maturity in making recommendations based on users’ past preferences (e.g., think: Netflix movie recommendations).
Microsoft’s ground-breaking work in the field of AI was the introduction of Co-Pilot earlier this year – in March 2023. It is one of the first AI assistants designed to work with people in a variety of different applications. It can provide a wide range of features for more productivity and efficiency. It is like having a brilliant research assistant that can get 80-90% of your work done. The technology is available in various Microsoft 365 applications, including Outlook, Teams, PowerPoint, and Word – and it’s nothing like Clippy. Co-pilot performs a variety of tasks: summarizing long emails, conversations, documents, and meetings; generating codes in a variety of programming languages using NLP from the developer (including updating the table structure to accompany new fields); etc. Microsoft plans to make the technology generally available to enterprise customers on November 1, 2023.
Jason Zander, Executive Vice President, Microsoft, recently showcased an example of how Co-pilot on their Teams application can get an employee up-to-speed on critical takeaways, action items, and next steps for a meeting he/she could not attend. Co-pilot has generated a meeting transcription, summary, and other critical elements of the recap:
Not only does the Co-pilot track all the tasks that came out of the meeting, but the employee can use natural language to interact with the Co-pilot for specific summaries, such as: Was there anything discussed in particular around creating marketing collateral for an event? What solutions were discussed? When is the deadline for submission? Co-pilot can successfully reference the transcript and highlight the relevant portions that answer these questions.
While the new AI technology sounds promising, security and privacy are of utmost importance. Microsoft is ensuring that sensitive customer data is safe and secure and compliance through safety standards. For instance, if you use Microsoft Azure OpenAI for Government agencies, it is accredited with FedRAMP High via Azure Commercial Cloud, CUI Support, and IL4/IL5.
Federal Government at Work
The potential application for AI continues to expand and evolve as the technology progresses and integrates with various industries and sectors. The Federal Government is already taking steps to explore its potential in enhancing its mission to serve its citizens in diverse ways:
- The Department of Homeland Security (DHS) recently announcedvi the formation of a new task force to investigate how Government agencies can leverage AI technology to enhance homeland security. In the spending bill for DHS, language is included that will fund AI and machine learning capabilities to review cargo shipments at the U.S. ports and for port inspections.
- The National Institutes of Health (NIH) exploredvii the use of AI in healthcare and researched generative models for drug discovery and medical image analysis. Reviews of existing literature on AI applications are already underway as dialogue around its application becomes more prominentviii.
- The Department of the Air (DAF) Force issued a broad agency announcement (BAA)ix on the development of next-generation AI capabilities to support the Department of Defence’s (DoD) Joint All Domain Command and Control (JADC2) initiative.
- The Departments of Energy and Defence have been leveragingx AI for technical projects to enhance precision and accomplish tasks beyond human capabilities.
Regulations and Considerations
AI is one of the most powerful technologies. The nation’s ability to leverage and take advantage of its potential will be critical to global and economic competitiveness. However, as with all things, enormous opportunities introduce proportionate risks. Actual AI adoption is still pending mitigation of those risks – particularly before everyday use across the Government.
In May 2023, Biden-Harris Administration announcedxi new actions to promote responsible American innovation through AI and protect citizen’s rights and safety. This step will ensure that the technology improves the lives of the American people and empower the Federal Government in its ongoing effort to advance its approach towards AI.
The Subcommittee on Cybersecurity, Information Technology, and Government Innovation discussed “How are Federal Agencies Harnessing Artificial Intelligence?”xii to survey the current and planned use of AI by its members. They also spoke on the associated risks with its widespread integration into various operations of the Federal agencies. During the session, Dr. Craig Martell, DoD Chief Digital and Artificial Intelligence Officer, emphasized the importance of responsible governance and accountability in using AI for national security while leveraging its advantages over adversaries. Dr. Arati Prabhakar, White House Office of Science and Technology Policy Director, reiterated the importance of reviewing all the work that has been done within the Federal agencies to harness AI capabilities. According to Dr. Prabhakar, if harnessed correctly, AI can create several opportunities for Federal agencies to strengthen value and better achieve their mission.
Initiatives are Already Underway to Achieve that Vision
Government agencies are evaluating some of the bills and legislative acts to embrace artificial intelligence within the U.S. Federal agencies as a whole or within individual departments, including:
- Artificial Intelligence for the American People Act of 2023xiii establishes a National AI initiative to coordinate the Government’s efforts in AI research and development.
- The AI Training Act of 2023xiv provides grants to support AI training programs for workers in the Federal Government.
- The AI Bill of Rightsxv provides a set of guidelines for the responsible design and use of artificial intelligence created by the White House of Science and Technology Policy (OSTP) to establish more regulations to govern AI.
- Enacted in 2021 as part of the National Defense Authorization Act (NDAA), the National Artificial Intelligence Initiative (NAII) by the National Artificial Intelligence Initiative Act of 2020 (NAIIA)xvi was established to ensure continued U.S. leadership in AI R&D, support development and use of trustworthy AI systems in public and private sectors, prepare the present and future US workforce for the integration of AI systems across all sectors of the economy and society; and coordinate ongoing AI activities across all Federal agencies, to ensure that each informs the work of the others.
- Treasury Board’s guidelinesxvii for generative AI was released in 2023 for employees who want to use artificial intelligence tools like ChatGPT at work to ensure that the technology is being used responsibly.
Government’s Way Forward with AI
Artificial Intelligence is not a future technology; it is already with us today and disrupting various industries. Harnessing the technology in a responsible and secure way will open enormous opportunities for private and Federal agencies to achieve their mission to serve the people of America. With Statista’s staggering projection of 181 zettabytes of data production and consumption by 2025xviii from digital engagement alone, AI provides a way forward for making sense of it all.
As adoption and use expands – particularly for the Federal Government – we need to be mindful of the risks it can bring and fix issues such as – cybersecurity threats, bias, privacy violations, and inaccuracy in data. Mitigating these along with proper governance is the need of the hour to have a long-term relationship with the technology.
As AI technology continues to evolve, there is a collective responsibility to define a clear relationship with the technology that not only enhances productivity and efficiency but also upholds the value of ethics, security, and fairness. With the right approach and ethical AI practices, the technology can become an invaluable asset for Federal agencies that thrive to better serve the citizens and navigate the complexities of the modern age.