The Center for Human-Compatible AI's goal is to develop the conceptual and technical wherewithal to reorient the general thrust of AI research towards provably beneficial systems. We are thus developing algorithms and conceptual tools to ensure the safety and reliability of AI systems. 

We think that the Trustworthy AI Assessment List is an excellent starting point for thinking about how to increase the trustworthiness of AI systems. We’d like to propose some improvements with reference to a domain in which we expect significant issues to arise: algorithmic trading. As AI technology advances, there will be strong incentives to develop AI systems which not only implement pre-specified trading strategies, but also learn novel strategies of their own. 

You will find our feedback in the attached document.