Writing the algorithm is the intent. Same as hiring a hitman.

IANAL, but I would think that it would be your (as the person deploying and maintaining this algorithm) responsibility to make sure this algorithm doesn't do anything illegal or disingenuous, either by putting safeguards into it or by constantly tracking it to make sure it doesn't do something bad.

You are legally responsible as if you had submitted the orders yourself.

I understand that... But as stated, the orders themselves are not illegal. It all hinges on intent of "your strategy to dupe other traders", but I fail to see how the algorithm creator would understand a black-box strategy developed by the algorithm. Eg let's say they algo was some form of reinforcement learning; it's not clear to me that you could even represent the policy space in a human-interperable way. All you have is inputs & outputs -- the AI has no "intent" beyond it's objective function -- say expectation maximization or risk minimization.

beam bot, im guessing the law would not be on your side.