Can I use AI to prove I'm human?

[purplepoppy]

Iwa hate doing those stupid CAPTCHA puzzles to prove I'm a human. So I was wondering if I could just get AI to do it for me.

[ScienceThumper]

Remember the guy, who was simping after @d_bone_steak's female cооmjar leftovers? His name had something with "muscle-". He argued something about the benefit of these puzzles, that they're more effective at blocking humans than blocking AI, but his reasoning was different. He thinks it's because most men don't squat nearly enough. ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ ﾠ

[JacobJordan]

No. The reason why is fairly simple. Ai is rigid humans are not. What this means is an ai lacks the flow of humanity. It moves along the x axis, y axis and z axis (possible) so it's movement's for a cursor are straight lines. Humans can't do that. Clicking the images is actually not the test. The test is how the mouse pointer moves. Along with human error of skipping the square with the 2 pixel portion of a light post or car or flower. An ai would select that square

[Jessica405]

I am just wondering how a touch screen copes? No mouse to move!

[Jessica405]

Like on the tablet I am using.

[JacobJordan]

@Jessica405 ai would tap in sequence left to right top to bottom. Humans typically do not. They tap based on what they see first so our tapping tends to be kinda random

[Jessica405]

It would be so easy to make it random!

[Jessica405]

Sure! Turn my tablet on its side or upside down.

[JacobJordan]

@Jessica405 still follows the same pattern. Plus screen orientation does not distinguish direction its sole task is to keep all information the direction and screen lock would just make it not auto rotate. Doesn't change which is "top"

[Jessica405]

The selection of which image to view first can be made random. AI is only needed to interpret that image. The first part is basic programming from high school IT.

[JacobJordan]

@Jessica405 well sure but humans can't even do random in our day to day lives. Programming true randomness is an impossibility. Even the human mind can't process or produce randomness. As such any ai by default is not random. Well none of the current ones at least. Nature produces true randomness. A genuine AI could in theory produce true randomness it would depend on several factors though

[Jessica405]

A simple RND function would be good enough.

[JacobJordan]

@Jessica405 still a pattern. True randomness includes clusters. And RND function or even an RNG is not truly random it follows a pattern.

[IslaTheWitch]

If you use credentials use "remember me" option and don't clean cookies.

If you use bound account don't clean cookies and don't use VPN...

AI still can fail miserably with images...

[mandyfire98]

There will be AI that will eventually just go around solving these and also trolling other AI.

[sawno]

Yes, https://chrome.google.com/webstore/detail/buster-captcha-solver-for/mpbjkejclgfgadiemmefgebjfooflfhl/related

[sawno]

Also special trick from me if the extention no longer works, click retry until you get one of those puzzles where you just got to click the right type of picture. Never complete these ones in your post since most people do them wrong so even if you get everything right the algorythm is completely off and its basically impossible.

So by refusing these terrible ones and only completing the sane ones you keep your statistics up that you are not a bot, and in doing so you avoid more of these terrible ones.

The second tip if the extention doesn't work is to do the audio challenge, those are WAY easier to complete.

[BlissfullyCurSed]

The logicality behind it is that human mind is flawed and does not work by following the same pattern all the time. The way an AI works is the complete opposite of the statement given above. It is almost always well-regulated and usually predictable. Maybe, that is possible, but I don't think chances are high that it will be successful everytime.

[hahahmm]

Those were a scam to get people to TRAIN AI to be more human, actually. Just like Apple's Siri was actually a plot to get people to help make AI's understanding of human language better. Same with voicemail to text.

[hahahmm]

So congrats, your job of teaching AI is finally over. One day a real life Terminator will thank you in person.

[91bryce91]

it won't matter. Boffins are working round the clock on ways to irritate the shit out of us online. And theyre very good.
DREAMING UP extra character needed for a password for example, phone verification code required before giving permission to fart, cancelling your account because you used a new browser you decided to try. etcetcetc

[Passerby2]

Lmao

[Red_Arrow]

So you want to use AI to prove you are not AI.

[Static_In_The_Attic]

👽 If you use AI to prove you are human I guarantee it will prove I’m not. 👽

[Sparkie460]

Before I answer, click on each box that has a face.

Yeah I hate doing those too

[rigmarole]

There have been automated captcha solvers sense captcha began.

[serious]

I don't think you can use AI to prove you are a human

[OddBeMe]

Funny majority of those are just distraction while they mine your search history.

[laly520]

some AI already able to do that

[Liketotry]

Not to me you can’t