Steering towards responsible AI – How world governments are addressing growth, ethics in AI.

Administration Issues Executive Order on Artificial Intelligence Development and Use

To expedite safeguards when it comes to artificial intelligence, the current administration issued an executive order on AI, which seeks to balance the needs of technology companies and consumer protections. With various tech companies in agreement with the safeguards set by the White House, the initial steps of this order are to assist in ensuring AI is trustworthy and helpful, rather than deceptive and destructive.

Some of these measures include having the National Institute of Standards and Technology (NIST) issuing standards for AI tools to be safe and secure. The U.S. Department of Commerce is assigned to issue guidance by labeling AI-generated content to help users differentiate between authentic interactions and those simulated by software. The order goes further into detailing consumer protections, worker’s rights, usage of scientific protections, respect for privacy and civil rights, or other oversight for rapid technologies such as ChatGPT.

In response to the executive order, the Cybersecurity & Infrastructure Security Agency (CISA) developed a roadmap to promote “the beneficial uses of AI to enhance cybersecurity capabilities, ensure AI systems are protected from cyber-based threats, and deter the malicious use of AI capabilities to threaten the critical infrastructure Americans rely on every day.”

Basic Applications of AI

Cyber divisions of organizations rely on AI for its capabilities to read into data, spot irregularities and signs of an attack, as well as keep data safe to the highest standard and best practices. AI is suited for tasks relating to analyzing huge amounts of data and providing insights into threats that allow for a quick, thorough response. When it comes to penetration testing, fraud detection, data-loss prevention, and other applications to be introduced, AI is a game-changer for businesses. Enterprises receive a significant edge in protecting and forecasting a brighter future through the application of AI. In a recent study by Microsoft’s Vijaya Kanaparti from Microsoft, Transformational Application of Artificial Intelligence and Machine Learning in Financial Technologies and Financial Services, “AI and ML models can effectively process large, unstructured, and high-dimensional data, enabling the extraction of valuable insights and the discovery of intricate patterns that may elude traditional econometric methods. The adoption of interpretable AI and ML techniques can facilitate a better understanding of the underlying mechanisms driving financial predictions and support more informed decision-making.”

Conversely, bad actors and cybercriminals have recently leveraged these technologies for their benefit. These tactics include an empowered approach to social engineering, deep learning algorithms that can figure out passwords, as well as using manipulation of visual and audio content to appear legitimate—also known as a deepfake. These actions can be applied for misinformation, extortion, harassment, and other harmful intentions.

Within the last few years, we have witnessed significant growth in the accessibility of AI systems and tools. To benefit from this technology, no longer does one need the skill and intelligence to operate a system, they can just pay someone who can do it for them.

Planning for 2024 Advancements in AI

From 2023 to 2025, CISA is aiming to adopt AI practices throughout the agency’s four-strategy goals. This includes implementing AI-based software for threat detection and mitigation, using AI systems to improve product resistance and risk awareness, and greater public awareness of threats and risks, as well as the responsible integration of AI across the agency’s infrastructure. 

Leveraging AI technologies for national security, CISA is partnering with the Joint Cyber Defense Collaborative to launch JCDC.AI– an effort to bring together AI providers and infrastructure operators to address risks. Not a website, although spelled like one, this initiative was established two years ago to unify efforts between public and private partners. Each year organizations partnered with the JCDC to “provide insight, expertise, and perspective to help identify our collective priorities for the coming year.”

To measure the impact of these efforts, the roadmap promises to achieve:

• Increased responsible use of AI software tools across internal workflows.
• Adherence to risk guidance and vulnerability management.
• Greater publications and engagements to support awareness of AI-related risks.
• A significant proportion of AI-focused guidelines and policy documents to be developed alongside United States inter-agencies and partners. 
• Heightened AI expertise in the workforce. Closing the skills gap to meet the new demand of AI professionals. 

The Administration and the various agencies operating are stressing a high priority on the education and mastery of AI technologies for cybersecurity. Understanding the legal, ethical, and policy aspects of AI-based software systems is a critical component that will be required from the workforce. The vision for the government’s approach toward artificial intelligence is to maximize benefits as well as protection from malicious use. Increased research and emphasis placed on training, awareness, and risk management will best equip the public and private sectors to meet the new challenges AI brings in 2024.