Salesforce Announces Public Sector Compliance Certifications, AI and Automation Tools to Accelerate Mission Impact

In order to avoid the siloing of best practices and lessons learned within each department, agencies should place a priority on publishing their efforts openly and communicating findings outside of usual intra-agency pathways. In addition to a technical focus on securing models, research attention should also focus on creating testing frameworks that can be shared with industry, government, and military AI system operators. In a similar manner to how automobiles are tested for safety, testing frameworks for the security of models can be established and used as a core component alongside the traditional testing methods used for vehicles, drones, weapon systems, and other systems that will adopt AI. From a technological standpoint, an additional difficulty is created by the fact that the field and technology itself is rapidly changing.

Governments today depend on AI algorithms to improve their decision-making as regards citizens. Sensitive data could find its way into the illegal hands or be used for iniquitous purposes. This calls for urgent government intervention in ensuring sufficient safekeeping of individuals’ private information. We are fostering industry support for SAIF with partners and customers, hosting SAIF workshops with practitioners and publishing AI security best practices. We partnered with Deloitte on a whitepaper on how organizations can use AI to address security challenges.

Where do AI regulations go from here ?

In addition, challenges that concern transparency and accountability are of importance in a government driven by AI. As AI systems grow into more complex and independent forms, individuals find it more difficult and worrisome to understand how and what their data is being used for, and if these algorithmic decisions remain fair. Governments need to proliferate mechanisms that support a transparent, accountable, and harm-free automated decision-making process. Google will continue to build and share Secure AI Framework resources, guidance, and tools, along with other best practices in AI application development. SAIF is our framework for creating a standardized and holistic approach to integrating security and privacy measures into ML-powered applications.

Second, the military’s unique domain necessitates the creation of similarly unique datasets and tools, both of which are likely to be shared within the military at-large. Because these datasets and systems will be expensive and difficult to create, there will be significant pressures to share them widely among different applications and branches. However, when multiple AI systems depend on this small set of shared assets, a single compromise of a dataset or system would expose all dependent systems to attack. A final avenue to poison a model is to simply replace a legitimate model with a poisoned one. Once trained, a model is just a file living within a computer, no different than an image or PDF document.

Data Privacy Hub

“By promoting accountability, data privacy, equitable solutions, and human review, we help our customers identify valuable use cases for generative AI and strike the right balance between human and AI,” said Gretchen Peri, managing director of Americas Public and Social Impact Industry at Slalom. Make the most of your live LEXI captions with LEXI Library – our easy-to-use cloud caption archiving and search tool. Via an intuitive web portal, LEXI Library allows you to view, search, edit and archive your live captions as raw text.

Taken together, these weaknesses explain why there are no perfect technical fixes for AI attacks. These vulnerabilities are not “bugs” that can be patched or corrected as is done with traditional cybersecurity vulnerabilities. From this understanding, we can now state the characteristics of the machine learning algorithms underpinning AI that make these systems vulnerable to attack. Third, the report contextualizes AI vulnerabilities within the larger cybersecurity landscape. It argues that AI attacks constitute a new vertical of attacks distinct in nature and required response from existing cybersecurity vulnerabilities. Given time, researchers may discover a technical silver bullet to some of these problems.

Using AI to Improve Security and Compliance

Commercial applications that are using AI to replace humans, such as self-driving cars and the Internet of Things, are putting vulnerable artificial intelligence technology onto our streets and into our homes. Segments of civil society are being monitored and oppressed with AI, and therefore have a vested interest in using AI attacks to fight against the systems being used against them. A secure cloud fabric is a powerful tool that can help the federal government meet its evolving data management and processing needs. It provides a secure and private multi-cloud connection that supports both data lakes and AI infrastructure.

It is necessary to hold those developing and deploying AI accountable to standards that protect against unlawful discrimination and abuse, including in the justice system and the Federal Government. Only then can Americans trust AI to advance civil rights, civil liberties, equity, and justice for all. Harmful consequences from AI systems can ensue for several reasons, even if neither the user nor the developer intends harm. First, it is difficult to precisely specify what we want deep learning-based AI models to do, and to ensure that they behave in line with those specifications. In other words, reliably controlling AI models’ behavior remains a largely unsolved technical problem.

Managed Cloud Infrastructure

(l)  The term “Federal law enforcement agency” has the meaning set forth in section 21(a) of Executive Order of May 25, 2022 (Advancing Effective, Accountable Policing and Criminal Justice Practices To Enhance Public Trust and Public Safety). (c)  The term “AI model” means a component of an information system that implements AI technology and uses computational, statistical, or machine-learning techniques to produce outputs from a given set of inputs. Some companies have countered this, saying they will face high compliance costs and disproportionate liability risks. The world has learned a number of painful lessons from the unencumbered and reckless enthusiasm with which technologies with serious vulnerabilities have been deployed. Social networks have been named as an aide to genocide in Myanmar and the instrument of democratic disruption in the world’s foremost democracy.

A central technology behind conversational AI is natural language processing (NLP), which enables machines to understand, interpret, and generate human language. Many U.S. agencies such as the Securities Exchange Commission, Internal Revenue Service, and the Department of the Treasury use algorithms to help direct enforcement resources for potential fraud cases. These programs are detailed further in Government by Algorithm, a 2020 report from ACUS. Our Enterprise license is ideal for government bodies who require multiple teams, as it leverages our unique Hub & Spoke architecture. Each team operates their GRC activities from a dedicated Spoke, ensuring data and operational separation with unrestricted access to modules, users, content, audits and a powerful AI engine, all connected to a central Hub for centralized administration, content management, and aggregate reporting.

Other AI Initiatives at the Department of State

Finally, local governments should regularly review their generative AI security policies to remain up-to-date and aligned with evolving security threats and best practices. Its features include a Content Library for turn-key compliance obligations and controls, an AI-enhanced controls builder, and actionable control task creation and linkages. The system simplifies evidence gathering for control effectiveness, and auto-maps controls to compliance needs, leveraging our AI engine and eliminating manual mapping. Integrated audit modules and a Trust Portal make auditing, sharing with stakeholders and proving compliance easy. Managing controls and policies is inefficient due to rapid regulatory changes, the labor-intensive process of developing and implementing high-quality controls, and the lack of a unified system.

Read more about Secure and Compliant AI for Governments here.