OpenAI recently came out with a paper titled "Practices for Governing Agentic AI Systems" introduced in the post at <a href="https://openai.com/research/practices-for-governing-agentic-ai-systems">https://openai.com/research/practices-for-governing-agentic-ai-systems</a>
It feels like it is meant to push forward the conversation on Agents and get many of the issues down in black and white. It is not a hard paper to read so I suggest you give it a shot if at all interested.
A couple of interesting points include the framework for describing AI agents and their "agenticness" including:
<ul>
<li>Goal complexity</li>
<li>Environmental complexity</li>
<li>Adaptability</li>
<li>Independent execution</li>
</ul>
And their "Practices for Keeping Agentic AI Systems Safe and Accountable" which include:
<ul>
<li>Evaluating Suitability for the Task</li>
<li>Constraining the Action-Space and Requiring Approval</li>
<li>Setting Agents’ Default Behaviors</li>
<li>Legibility of Agent Activity (note they don't say explainability)</li>
<li>Automatic Monitoring</li>
<li>Attributability</li>
<li>Interruptibility and Maintaining Control</li>
</ul>
All in all a good overview that highlights key issues. There will be a lot of research in this area and the blog post mentions a new grant program for studying agentic AI systems practices.

OpenAI recently came out with a paper title "Practices for Governing Agentic AI Systems" ... but there are more than three laws.

The Three Laws of Agentic AI Systems