Who to sue when a robot loses your fortune

(Kenny Grant) #1

https://www.bloomberg.com/news/articles/2019-05-06/who-to-sue-when-a-robot-loses-your-fortune

4 Likes
(Chris) #2

I doubt this will be the last. Interesting dilemma though, courts and regulators have been slow to move on this issue so they might have their hand forced here. I suspect the US will move quickly on this, the SEC will not want to be blindsided.

1 Like
(David Kent) #3

Interesting read. The issue comes down to completely trusting a robot to be the brain. I like using tools for guidance rather than execution.

I’m sure one day we’ll get there but I doubt it’ll be with this iteration of AI.

It sounds like the AI was still owned by the selling firm as overdue fees were brought into question. If it was purchasing the software and then walking away I’m sure it’d be treated differently. Feels no different from people taking their IFA to court over poor performance or misleading risks.

(Louis Otto) #4

I guess it really depends on how it was sold. If the salesperson guaranteed profits (with or without knowing the full workings of the system he is selling, or explaining the fees), then sure, he is at fault and deserves to lose the case. If, however, he did his job and stated the quite frankly cookie cutter statement of “investment value can rise and fall, and past performance isn’t an indicator of future performance” he’s pretty much covered, right?

4 Likes
#5

It does sound as if it being a machine (or not) is a secondary factor, and how it was sold or not sold the key thing. After all, the actively stock picking humans experts at an investment manager would also be somewhat black boxy from the retail investor’s perspective.

(See also the conversation about Marketsflow! - are they saying “our clever AI might do better than other strategies” or “it’ll return 20-25% annually”?)

(Kenny Grant) #6

I think the key to not being sued is to say

it has returned 25% annually… before fees, your investment may go up instead of down etc…

Personally I don’t think this suit has merit unless there was miss-selling, but it’s an interesting question - with many complex neural net approaches humans simply do not know why it does what it does.

This is very different from a traditional manager who has all sorts of (often misleading, spurious, or wrong) explanations for trades and losses or gains.