The Executive Order issued by the White House on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence marks a significant step towards establishing a robust framework for AI governance.¹ As a compliance expert, I recognize the critical importance of this directive in setting clear guidelines and expectations for the responsible development and deployment of AI technologies.
Here is my take
1. Secure and Responsible AI Development
The order underlines a simple fact – we need strict rules and safety measures in place when creating and rolling out large-scale AI models. Why? Because it’s the only way to defend against the risks and flaws that can mess up the workings and security of AI systems. So, the question isn’t whether we need thorough testing and evaluation methods. The question is – how quickly can we establish them to safeguard AI applications in every field?
2. Safeguarding Sensitive Information
The executive order puts a strong focus on the secure handling of sensitive information. It restricts access to high-value computing resources and cutting-edge AI models for foreign entities. This step ensures safety against potential misuse or exploitation of our critical information and technology. The government shows a clear commitment to protecting national security and intellectual property through this action. It’s a strategic move to retain our competitive advantage in AI.
3. Accountability through Identifiable Markers
Use identifiable markers to keep track of AI-generated content. This innovative method helps us trace where AI content comes from and keeps accountability. With markers, we can verify information’s authenticity and make sure we’re using AI systems in a way that’s both ethical and responsible.
4. Enhancing Data Privacy and Security
The executive order demands the creation and use of technologies that bolster data privacy and security for a range of applications and platforms. Taking a proactive role in data protection is vital for establishing trust in AI systems. It guarantees the safeguarding of users’ privacy and ensures ethical, responsible data usage.
5. Preventing Biases and Ensuring Equity
Taking active steps to stop bias and discrimination is at the heart of creating fair AI systems. This effort helps encourage inclusivity and diversity. AI technologies and tools should be made for everyone and benefit every part of society.
6. Safeguarding Employment Opportunities
The executive order explicitly calls for putting policies or measures in place to protect jobs as AI continues to evolve rapidly. Recognizing the potential for job displacement and the need for a skilled, AI-ready workforce is vital. We need to equip our workforce with the skills and knowledge they need to excel in an AI-centric future.
What it means
While the detailed regulations for AI safety and security following President Biden’s Executive order have yet to be created, best practices aligning with the NIST AI Risk Management Framework are actionable now and mitigate both compliance and reputational future costs.² I recommend companies manage their AI strategies & investments by following the below:
- Align your AI development with the guidance found in NIST’s AI Risk Management Framework.
- If you are buying AI, ensure your vendors align with the NIST AI Risk Management Framework.
- Standardized risk assessments soon will contain questions surrounding AI. Be prepared to issue and receive hundreds (or more) assessments on how AI is used and managed in your business.
- Laws and regulations will likely require testing, usage, and disclosure of how AI is used. Many customers will require their vendors to adopt it. This makes AI monitoring and tracking in the workplace as crucial of a risk management activity as compliance document versioning, tracking, and attestation. For this, you should adopt an AI orchestration platform. AI Orchestration platforms deploy standard and custom AI models quickly into your business. They provide an audit trail of how and when AI is used as well as what version of AI is used. This understanding will become a common question for business managers, customers, lawyers, regulators, and other stakeholders. Starting your AI journey with an expandable AI orchestration platform will provide you with a competitive advantage, cost savings, and third-party preparedness that you will need in the future to comply with the AI executive order.
Sources:
Read our comprehensive LLM testing guide
See how several LLMs fared in our test and steps to run your own tests for each of your use cases.
