Artificial intelligence promises to bring a new era of productivity and efficiency to government services and activities. Agencies across the US Federal government have been investigating, planning and implementing AI-based systems that take advantage of advances such as large language models, generative AI, and novel uses of machine learning. The new administration has begun discussing an “AI-first” strategy to streamline government services and decision-making.  A new Request for Information from the National Science Foundation “requests input from all interested parties on the Development of an Artificial Intelligence (AI) Action Plan.” This action plan will likely result in recommendations on various topics, including applications of AI in government and public services, cybersecurity, data privacy, and the effective and practical requirements for information governance as the action plan takes shape.

Over the last year, US government departments and agencies have been planning and developing innovative AI use cases. Each agency has published use case inventories. For example, AI use cases are under development at NARA, the Department of Homeland Security, the Department of Health and Human Services, and the General Services Administration. The variety and scope of these AI use cases are also vast. 

Central to successfully implementing AI is the need for great content management and, with it, content-aware AI systems that can safely and securely deliver trusted machine learning and generative AI capabilities. When combining these elements with AI governance principles, IT and business analysts can be assured that they build on a strong foundation to deliver AI-first productivity and efficiency within each agency.

High-impact opportunities for AI in government

Within each agency AI use case inventory, some common use cases stand out:

• Intelligent document capture and workflow processing. AI can evaluate incoming documents, images, and faxes to recognize text and handwriting, extract metadata, classify content for downstream processing and apply security policy. This can help streamline thousands of individual government processes. 
• Case, contract, and project management applications. Invariably, even highly predictable processes such as law enforcement or legal case management, contract and project management involve large volumes of complex, unstructured documents. Generative AI search and summarization can help users navigate to essential documents quickly and answer questions in minutes, potentially saving hours of difficult document review and research. Moreover, when combined with a content management system, the everyday management of information organizes and governs the content; it also facilitates AI context, significantly improving the accuracy, trustworthiness, and security of AI within the workforce.
• Research. Research databases invariably provide text and field searching. However, this sometimes requires users to review dozens or potentially hundreds of documents to find the correct answer. Generative AI that can connect to many different research repositories can consolidate the search process and help speed the research and citation process. AI systems can also provide additional classification and scoring capabilities to evaluate research effectiveness and validity.
• Human resources and recruiting efforts. Human resources processes involve large volumes of unstructured and semi-structured content. Generative AI can help to compress workloads when dealing with applicant pools or querying for expertise or field experience within the department.
• Freedom of Information Act support. Responding to a FOIA request can be time-consuming and expensive for an agency. Moreover, assuring coverage of complex topics across multiple repositories can complicate the process. AI systems that can access numerous repositories using AI-trained natural language processing (NLP) and provide content labeling and redaction can help streamline these processes and facilitate FOIA responses efficiently.

Use cases that test limits

Some use cases test the limits of traditional AI and machine-learning systems. This is especially true when dealing with complex file formats or rich media. AI systems need to establish secure access to information, the ability to work with many different file formats, and appropriate AI for that media.  As an example, some use cases that would test the limits of more traditional systems include:

• Speech-to-text transcription and speaker recognition.  The ability to turn speech into text accurately, translate it to a common user language, and identify the speaker is called out in several inventories.
• Machine vision and object detection. Various use cases require the ability to detect objects and conditions from video feeds. Surveillance cameras, weather cameras, and other video feeds can provide valuable source data, but reviewing the content can be laborious and error-prone without AI. AI can speed up these detection and review tasks and even increase accuracy.
• CAD and Engineering management. CAD drawings are used for knowledge management, reference, and research applications. Content management that can handle engineering use cases and incorporates generative AI can help identify drawings for collaboration and reference and quickly cite the actual drawings.  

OpenText™ Knowledge Discovery is the foundation of a comprehensive AI strategy

OpenText™ Knowledge Discovery provides a complete solution for addressing complex or large-scale AI use cases for government agencies. With powerful, built-in full-text search, a generative AI-based natural language interface, and visualizations illuminating your data's hidden relationships, it is the perfect tool for ad hoc search and more directed Q&A applications.

AI content management helps organizations and agencies understand their content and achieve productivity by identifying content quickly, labeling and protecting it, and intelligently putting it to work.

Some of its many capabilities include:

• Real-time categorization and machine-trainable classification can instantly group and direct content to key processes. 
• Review, workflow, and redaction capabilities help facilitate collaborative review of vast content stores, label content, initiate workflows and secure and protect content.
• Metadata enrichment can identify sensitive or privacy-related information to apply critical access controls and security labels.
• Rich media AI allows agencies to generate audio transcriptions and translations, identify speakers, and provide facial and object recognition in images and video.

Notably, OpenText Knowledge Discovery can connect to existing content repositories (over 160 out of the box) and process over 2,000 file formats. With over 20 years and dozens of patents, OpenText Knowledge Discovery is a comprehensive, secure and scalable solution for addressing AI-first government.  

By integrating great content management with generative AI and machine learning technologies, hundreds of high-value, highly productive AI use cases can be quickly implemented. Use cases in various domains, such as case management, research, human resources, and more, showcase AI's transformative potential to improve government efficiency and productivity while remaining safe and secure. By leveraging advanced AI capabilities, organizations can streamline complex processes, manage vast amounts of unstructured data, and improve decision-making.

The post AI-first government productivity and efficiency appeared first on OpenText Blogs.