Regulation, Safety, and What Comes Next

Reinforce AI safety: adopt stringent testing, uphold clean data practices, and leverage open-source safety frameworks for reliable, responsible AI systems

Feb 11, 2026 9m 27s

Elizabeth Seger

Expert Insights

Elizabeth Seger underscores the imperative for responsibility and safety in AI-assisted software development, warning that current practices and regulations are insufficient. She highlights the potential for troubling misuses of AI technology, particularly in deepfakes and content moderation, but also points to opportunities in AI safety testing and the emergence of AI assurance industries.

Elizabeth provides a comprehensive view that calls for stringent testing, careful data handling, and clear regulation for AI systems.

Hear her outline:

Why today’s AI regulatory environment remains relatively loose, leaving much of the responsibility for safe deployment in the hands of developers.
How ethical risks—especially around misuse of AI-generated content—make robust moderation and safeguards non-negotiable.
What emerging open-source safety frameworks like AI Verify and Roost signal about AI becoming embedded in everyday life?
How the rise of an AI assurance industry enables companies to validate model safety and performance through independent expertise.
Why transparency, rigorous testing, and accountable data practices are foundational to building trustworthy AI systems.

AI-assisted software is racing past classic determinism, causing us to rethink the entire development process. It requires a new language and system, not a traditional one that plasters over Agile just to keep it alive. Pristine data is crucial; if we feed our models with bad data, we get worse than bad results, we get misleading ones. Complexities can mount up, of course, but adopting tools like OpenAI's Gym can help prioritize safety in AI environments, boosting reliability and user trust. We're on the cusp of AI's golden age, but we need a solid foundation to ensure a smooth transition.

Monterail Team Analysis

Here's how to manage responsibility and safety in AI-software development and navigate the evolving regulatory landscape:

Be Responsible: As AI developers, it is vital to prioritize safety and quality. Test your models thoroughly and ensure the data you're using is free from harmful content.
Understand the Legal Landscape: Get familiar with current AI regulations and the ones applicable to your products. Monitor policy changes at both the state and federal levels, as these can impact your products and services.
Assess Risks and Ethical Implications: Training an AI model brings about ethical and legal considerations, especially over misuses of AI-generated content. Be proactive in mitigating potential risks and navigating ethical grey areas in AI adoption.
Explore Open Source Safety Frameworks: Use open-source safety frameworks and tools like AI Verify and Roost to ensure your AI models are safe and reliable.
Consider AI Assurance: If you don’t possess internal expertise, consider partnering with AI assurance services. These companies can provide testing and maintenance of your AI systems, ensuring their safety and effectiveness.
Prepare for the Emergence of the AI Safety Industry: With the growing reliance on AI technologies, an AI safety industry is expected to emerge. Plan your AI development pathways in line with this emerging reality.
Stay Self-Regulatory: Look beyond expecting regulations from the government, test, clean, and ensure your AI model is safe and high quality. It's not just about being a responsible AI developer but also about protecting your business and stakeholders.