Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Hacked billboards make Teslas see phantom objects causing them to swerve or stop

newsweek.com

Post reply on HN
Print view View original

71–80 of 264 posts

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#71

Post by dragonwriter »

Buttons840 wrote: post #59
Polylactic_acid wrote: post #24

Earlier quoted context omitted.

The wonderful thing about IoT is the criminal can be on the other side of the world with no chance of any justice coming to them.

How long until someone claims one of these images designed to cause trouble for AIs is free speech? And are they right?

> How long until someone claims one of these images designed to cause trouble for AIs is free speech? And are they right?

They are not, any more running someone through a printing press to kill them is “free press”. Regulating intentional harm without reference to the particular means is not a violation of free speech or other expressive rights just because someone uses a tool that could instead be used as an expressive medium as a mechanism of inflicting the harm.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#72

Post by dragonwriter »

someguy101010 wrote: post #65
Buttons840 wrote: post #59

Earlier quoted context omitted.

How long until someone claims one of these images designed to cause trouble for AIs is free speech? And are they right?

Id say no to the free speech since that sounds similar to shouting fire in a movie theater

Shouting fire in a movie theater is not a good example of unprotected speech, and as such it's not a good place to start reasoning from in deciding of other speech is unprotected. Otherwise, you start banning core political speech like advocacy against the draft, just like the since-overturned case that bad example comes from.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#73

Post by couchand »

dwohnitmok wrote: post #44
brianwawok wrote: post #36

Earlier quoted context omitted.

But it would fall for a cardboard stop sign on a stick held up by a troublemaker.

As would a human given a sufficiently official-looking stop sign. In fact I think it's actually impossible for an automated car or human without figuring out the "troublemaker"'s identity whether the stop sign is fake or not. And it's not always possible to pin down someone's identity (if the person is wearing a brightly colored construction vest what do you do?).

There's a lot of construction in our neighborhood right now which made me start thinking about that possibility. It kind of sounds like something out of a heist movie (or taken to the extreme, Die Hard With a Vengance?), but really, who would know and care enough to ask the questions to notice something like that?

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#75

Post by arminiusreturns »

smallstepforman wrote: post #35

There are many reasons why I dont believe in self driving cars, the first (and strongest) is to imagine that the software was written by my collegues (and myself).

Having been in the industry, you have no idea.

I never rode in our vehicles... for a reason.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#76

Post by maxkwallace »

This is a special case of a general class of vulnerabilities in AI models, where an adversary can cause undesired output from the model by constructing input data not represented in the training set. However, it is legit much more concerning than, e.g. the issue of image classification models mis-identifying well-constructed noise as "panda".

This is currently a research frontier for AI so us non-experts likely won't be able to say a ton about it.

I thought this was a good talk on the issue: https://www.youtube.com/watch?v=SS9DMr4VkbY

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#77

Post by tshaddox »

someguy101010 wrote: post #65
Buttons840 wrote: post #59

Earlier quoted context omitted.

How long until someone claims one of these images designed to cause trouble for AIs is free speech? And are they right?

Id say no to the free speech since that sounds similar to shouting fire in a movie theater

And yet you don't have to look far from here to see lots of people claiming that a social network is violating freedom of speech if it censors or in any way limits the distribution of false claims that are intended to and likely to cause people to take dangerous actions.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#78

Post by TaylorAlexander »

Polylactic_acid wrote: post #24
mattbee wrote: post #20

This is clever, and Tesla's software is awful, but anyone doing this for real, with intent to divert cars from the road etc. would be committing a pretty serious crime.

The wonderful thing about IoT is the criminal can be on the other side of the world with no chance of any justice coming to them.

The trick is to construct a real looking advertisement that “accidentally” triggers this effect, and order the ad placed on digital billboards in Silicon Valley through a fake company located somewhere far far away.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#79

Post by tshaddox »

dragonwriter wrote: post #72
someguy101010 wrote: post #65

Earlier quoted context omitted.

Id say no to the free speech since that sounds similar to shouting fire in a movie theater

Shouting fire in a movie theater is not a good example of unprotected speech, and as such it's not a good place to start reasoning from in deciding of other speech is unprotected. Otherwise, you start banning core political speech like advocacy against the draft, just like the since-overturned case that bad example comes from.

The fact that that example became well-known due to the awful decision in Schenck v. United States does not mean that it's not a good example of speech that ought not be protected. Speech that is false and intentionally causes a panic should not be protected.

Re: Hacked billboards make Teslas see phantom objects causing them to swerve or stop

#80

Post by userbinator »

TaylorAlexander wrote: post #78
Polylactic_acid wrote: post #24

Earlier quoted context omitted.

The wonderful thing about IoT is the criminal can be on the other side of the world with no chance of any justice coming to them.

The trick is to construct a real looking advertisement that “accidentally” triggers this effect, and order the ad placed on digital billboards in Silicon Valley through a fake company located somewhere far far away.

There have already been (audio) ads that trigger speech-recognising machines like Alexa, and in that case I would very much say it is free speech.
Post reply on HN