Microsoft knows a little something about captive audiences. Witness, for example, this patent application for a CAPTCHA system where you prove you’re a human by recognizing some awesome products.

adhip.jpg

Microsoft says that this innovation creates “an effective way for advertisers to deliver their message to a captive audience. Unlike so much web-based advertising that accompanies popular web portals such as search and news sites that users can easily ignore, here the user must actively engage in reading and understanding the content in the advertisement in the HIP (human interactive proof) challenge in order to identify the solution to the challenge.”

I look forward to the widespread adoption of this technique. I am a human, after all, and I can recognize any product you got. Hell, I’ll probably buy one!

Via TechFlash.

Anthony Hecht is The Stranger's Chief Technology Officer. He owns no monkeys.

12 replies on “I Am Human, For I Have Read Your Advertisement”

  1. This is a trick question. Is the answer XBOX or XBOX 360?

    And this won’t stop the spammers. Text recognition software can be trained to read logos, and it would be against the Microsoft corporate brand to sufficiently distort their logos to thwart spammers. So, I don’t forsee this as a actual viable alternative to the normal CAPTCHA images you see now.

  2. @6 I don’t think the intent is to actually be better than current captcha systems, I think the point is to convince advertisers and website operators it is for the purposes of generating revenue.

  3. The requirement for use of a web form is that you are human, not that you have read an advertisement then passed a comprehension test.

    Requiring a confirmed ad impression just to complete a purchase is a greedy, double-dipping abuse. I will boycott any retailer that adopts this — even you Amazon, my love.

    Now if that ad revenue was passed on to the consumer, as an option to earn savings, “Save $.xx by correctly stating what brand of detergent you see above.” That would be a fair implementation of this patent.

  4. Seems like this would be a lot easier to get around. You are working with a smaller dictionary of words and in the example the words are not smudged at all. The problem becomes find as much text as you can in the image, then find the closest matching product name to a word in the text. (Since we are not skilled enough to read image text exactly.)

Comments are closed.