Hacker News Reader

Comment by rfugger | Hacker News Reader

You sue whoever misrepresented the robot's capabilities to you. There's nothing wrong with making a robot that's bad at trading. There is something wrong with pretending it's not bad at trading when you're selling it.

View on HN

Your answers diverge, depending on who is operating the "robot". If it's being presented as a service by an investment firm, then you obviously have a standard dispute with the investment firm, the same as if one of the human brokers went nuts. If was sold as a software system to be self-administered, then you have an issue with a vendor. The latter may hinge solely on misrepresentation, because software is generally sold with little warranty for precisely these reasons.

So much money slushing around can make things a little grayer because in either case the investor didn't actually perform any work themselves. But I'm confident they can use some of that money to discern who ultimately owned what, who employed who, and what type of relationship was created.

I feel like newswriters just love to generate hot air with this "sue a robot" trope. Who to sue for that?

Is it even possible to misrepresent the robot's capabilities? In the UK most (all?) advertisements come with cautionary notices such as "capital at risk" or "past performance not indicative of future returns" or similar...

Although obviously such warnings can't reach some people... My favourite example of this was the collapse of OptionSellers.com following the huge move in Natural Gas futures. I mean, they were selling options - it's right there in the name...

https://www.bloomberg.com/news/articles/2018-11-19/hedge-fun...

IANAL but I could see a lawsuit being possible if historical "actual" performance figures were misrepresented or back-tested theoreticals (read: overfit) were passed as actuals.

What if it good at trading but can be unlucky?

If it's a robot, designed, programmed appropriately, and trained using the correct data to digest and understand later data, can it be unlucky?

John Maynard Keynes' remark that 'the market can remain irrational longer than you can remain solvent', springs immediately to mind.

Yes, unlucky just means that random/arbitrary decisions or events didn't "favour" you.

Then I'd say the probabilistic nature of the model was not represented correctly by the seller.

Saying the algorithm makes 95% winning trades may not mean much if those remaining 5% of trades cause you to lose most of your value.

Protip, do not try this argument in Vegas.

Good point, but at least the probabilities of casino games are generally known and available AFAIK. People just tend to not care or have a confidence bias.

What if that person didn't fully understand how the robot worked? The best your could hope for is negligence

If they didn't fully understand how it worked, but fraudulently represented that they did, they'd be on the hook. It all hinges on what he was promised when he bought it. If they had a good disclaimer, then he might be out of luck.

There are places where adopting AI is going to be difficult for this reason - notably anywhere that has a legal requirement to explain why a decision went a particular way. That includes things like insurance risk models, pension investments, etc.

In these cases, I wonder if it would cause a movement toward more interpretative models with more mathematical backing.

For example, if not being able to explain a model leaves an organization open to litigation, they may instead rely on statistical based learning methods even if they performed less well.

The losses made by software training may be the least problem for the institution.If case goes to court,they may be required to reveal the reasoning behind software's decision and that means going through the code,as all that marketing bs called AI is just a trading algorithm trained on some datasets. No company would want to reveal the code.

What if some other person really understands how the robot worked and half of the money you lost ended up directly in their accounts?

Is cheesing an algorithm illegal?

Reply on news.ycombinator.com