Remarks on the Trump Administrations *so-called 'AI Action Plan'

The following was written without CHATGPT, etc.

Hello

There will inevitably be overlaps between ‘so-called’ AI and robotics. This warrants, I believe, my offering periodic commentary on ‘so-called’ AI. 

The Trump Administration recently put forward an action plan regarding this domain of innovation. *And I habitually use the term 'so-called' whenever I have to reference the term 'AI' - given my aversion to it.

While I agreed with most of what was said (and admit I was glad to see that the term ‘AI’ came under scrutiny), I do disagree majorly with at least one premise of this action plan - in particular, it was suggested that in order for the US to maintain dominance with the so-called ‘AI’ tool, we should not allow for individual States in our country to create their own standards around it. To repeat, I disagree with this majorly. 

We SHOULD absolutely allow for States in the US to develop their own codes of conduct and implementations for these tools. For certain, one of the few benefits of having many States in our vast country is that we can, on the whole, check and balance one another - and negotiate difficult, fringe matters where there's no known, non-religious best practice.

I heard our American State system once compared to the scientific method - in that with our State system, we can test out numerous methodologies on difficult fringe matters and, after a time, begin to gain a better understanding of what works and what does not. It’s hard to imagine anything being more difficult to understand than this technology. Otherwise, it seems rather dangerous to try and attempt to create sweeping, legally binding standards for a phenomenon of this unknown variety.

To address the point of concern directly, “how does the US stay on top of this technology without centralizing power / standardizing solutions?” I suppose my initial rebuttal would be we don’t know yet what constitutes a good standard for these technologies - and the best possible solution, it seems to me, will not just be one that makes the most efficient possible tool, but one that builds a robust, human-centric culture around these tools. 

We know that, in theory, given the time and resources, we can make whatever tool we want and to whatever specifications we wish it to have. But we haven’t got a clue, yet, on how we might actually live with these tools. 

Thank you for your consideration,

Michael Christensen-