Skip to main content

Is there a way to make the images in ReCaptcha V2 accessible by keyboard?

reCAPTCHA v2 challenges should be accessible by keyboard. What issue do you see?


As a neurodivergent user (Asperger), I have always faced severe accessibility barrier with reCAPTCHA v2 image challenges, consistently failing with an estimated 80%+ failure rate. It´s gotten so tfrustrating to be classified as a robot, and sometimes I simply cant pass, even when I guess what the expected answer vs the accurate one is

The root cause is the conflict between literal interpretation and the system's expected non-literal answer (i.e., the system is trained by crowd consensus, not precise definition).

Example: When asked to select a 'traffic light,' I select all tiles containing any part of the light(including the pole and cables). The system rejects this, I assume it is expecting the non-literal answer of only the light housing. (But even when I guess at the expected answer, I still have trouble.)

This design flaw punishes literal interpretation, wrongly classifying a legitimate human user as a bot.

Any chance you can consider this? Or have you? I am just one person, and as an Aspy, dont have much contact to many people but imagine this wil be very common for people on the autsim spectrum.

PS: Given the high prevalence (305) of people with Aspergers in the IT field, perhaps you can simply collaborate with or pay special attention to what the feedback is from your neurodivergent/Asperger's colleagues or community groups to audit challenge design for cognitive accessibility and precision.

Thanks