This captcha has gone viral, receiving mass attention on Twitter and Tumblr.
Image Courtesy of Twitter
The phrase is a bit silly, but it accurately captures the feelings of frustration that often come with trying to solve a captcha. It’s a lot like this–
Gif courtesy of Teendotcom.Tumblr
And it’s not just white girls who are having issues with captchas. According to a study conducted by Stanford Theory Group, a section of Stanford University, this ability to ‘even’ is a problem experienced by all types of Internet users. The study found that across a variety of users, image captchas had “an average solving time of 9.8 seconds and three-person agreement of 71.0%, and audio captchas being much harder, with an average solving time of 28.4 seconds, and three-person agreement of 31.2%.” And those statistics are from 2010, when captchas were as easy as this–
Image courtesy of Developers.Google
In 2014, as bots have gotten smarter, the captchas we encounter are much harder. It’s no surprise that four years later many Internet users still can’t even.
We know that not all online advertisements are watched by users (advertisers themselves probably don’t even watch them). But did you know that some ads are still not even viewable? As reported by MediaPost, despite viewability gains being made by the advertising industry, the issue has not gone away. Its a lot like plucking off the head off of a dandelion, or in our case ad fraud. When we forget to pull up the weeds/hackers as well, the problem is just going to grow back in a more creative, and likely problematic way.
Photo courtesy of WoollyGreen
According to Kevin Lenane, ad viewability is plagued by a number of issues–multiple videos appearing per page, video partial appearance, inconsistencies with video autoplay and audibility. Engaging with online ads is already a limited sensory experience, and by eliminating sight and sound, users are left abandoned by advertisements. It’s not like we can taste an ad, even though that would be awesome.
The larger problem though, is that most often neither advertisers nor users know that any of this is happening. As MediaPost prompts, it’s time for the industry to try something new–knowledge. We need to know the issues of viewability “BEFORE the campaign runs at the media plan stage, DURING the campaign as a blacklisting/whitelisting opportunity, and AFTER the campaign as an analytics and optimization check.” It’s time to pull some weeds.
Are computers becoming more human? If this week’s press is any indication, the answer would seem to be yes.
As reported by NPR, for the first time ever a computer has successfully passed the Turing Test, an industry’s benchmark for accrediting intellectual thought to machines. Although the bot, named Eugene Goostman, was only able to persuade a third of the judges of its humanness, and valid scrutiny surrounds Goostman’s true capabilities, the achievement nonetheless demonstrates advancements for computer intelligence.
And bots aren’t only pretending to be humans, they are beginning to perform human jobs as well. The Associated Press (AP) announced that it will use story-writing software to produce the tedious corporate earnings reports for non-high-interest companies, and bots are also being used to keep World Cup venues safe. The question remains, though: are computers actually becoming more human, or just getting really really good at pretending?
NPR seems to believe the later, and we agree. As the publication writes, “it’s one thing to be able to express emotions and another to really feel them.” While computers may now be able to feign a range of human emotions and accomplish a variety of previously human tasks, do they really feel anything in the process? Can a computer experience the frustration that prompted AP to push monotonous tasks like earning reports to bots in the first place? Can a computer celebrate a win at the World Cup while sweeping for suspicious packages? Probably not.
This fascination with computer-human relationships is a hot topic across many fields. As NPR notes, we’re dedicating entire movie plots to it. 2013’s Her best explores this complicated advancement.
Photo courtesy of ThinkProgress
While Theodore might just be an overly paranoid man (I mean, look at that brooding), and Samantha an exceptionally perceptive bot, the disconnect between the two demonstrates real world implications about the future of computer and human interactions.
As NPR duly notes, “what makes a computer seem human isn’t how we perceive its intellect but its affect.” Scarlett Johansson (Samantha) might portray a damn good human, but as with bot writers and bot police officers, computers may not be be able to interact with us in truly genuine ways. Theodore’s frustrations were felt by many who saw limitations in Eugene Goostman’s passing of the Turing Test. “Computers don’t even go about making small talk the same way we [humans] do,” but genuine believability is a highly subjective experience.
Turing himself brings up an interesting point about determining genuineness: “how can you tell? After all, how can we know for sure that anybody else is really conscious, except by a leap of faith?” At Are You a Human we’ve chosen to put our faith in humans first. Although we are constantly allotting this benefit of doubt to each other, we aren’t reading to extend that courtesy to computers just yet.
Right now, the issues of viewability and ad fraud are being treated like a bad rash that just won’t clear from the face of digital advertisements. They’re annoying, they’re ugly, and rather than preventing them we often find ourselves scrambling for solutions after they’ve already spread. While the Media Rating Council has been working to improve low industry standards, a recent blog from MediaPost offers three ideas of its own to fix these blemishes.
Image courtesy of RojerThat
Impose Real Penalties: “Advertisers and agencies should implement zero-tolerance policies and shut off services and networks completely for all future business opportunities when they have experienced fraud…Demand full refunds on campaigns, not just for the fraudulent portions. Recover fees from agencies.”
Remove Artificial Incentives: “One of the biggest drivers and ‘enabling conditions’ of robotic traffic fraud is media buyers’ desire to push overall costs-per-thousand down. As easy way to cut CPMs in half is to use robots to double a site’s or network’s traffic. We need to focus more on cost-per-business-objective, not just on cheapening intermediate metrics.
Remove Automated Site Sign-Ups: “Interact with real people, always. networks and exchanges make it easy and automated for almost anyone to sign up and get ad tags… How about insuring that there are real people behind sites.”
As Dave Morgan, CEO of Simulmedia, notes in the MediaPost blog, “clearly, there are a lot of folks focused on solutions.” And solutions like these are good, but penalties and removals lack some of the forward thinking necessary to sustain long term industry improvements. We need to be more proactive about targeting the sources of ad fraud as well as their effects, all the while ensuring that every ad is really reaching a real human. Proactive Acne Treatment is called proactive for a reason, and it’s time for the ad industry to adopt some of that dermatological urgency.