Is it Google's job to fix society?
- 13 July, 2015 20:12
Social problems exist. Sexism, social stigmatization, crime and other problems have always plagued humanity.
Increasingly, however, organizations are calling on Google to solve or help solve these social problems, usually at the expense of the quality of the company's own products.
Here are a few examples.
Carnegie Mellon University and the International Computer Science Institute created software called AdFisher, which analyzes Google ad targeting, including job postings. They found that male users were shown ads for high-paying executive jobs more often than women were.
The decisions about which ads to show to which users are based on two sets of data. One is the user information that Google collects. The other is the advertiser's data, which is a combination of data that the company harvested and hand-entered information about demographics.
Google's job is to faithfully take the user signals and the advertiser signals and enable communication between the two parties.
It's possible that sexism at Google determined or contributed to a sexist outcome for the targeting of the executive job postings. To the extent that that is true, Google must correct the problem as soon as possible.
However, it's more likely that Google's algorithm is faithfully conveying the sexism that in fact exists in society.
If that's true, what should Google do? Should it continue to offer "neutral" algorithms that deliver whatever biases and prejudices are contained in the inputs? Or should it counter society's biases and program corrective social engineering into its algorithms?
In other words, should Google take the inputs that skew toward sexism and deliberately weight them in the other direction to produce outputs that create the illusion that sexism doesn't exist?
The Internet has radically increased the scope and consequences of social stigma, or at least people believe it has (reading Nathaniel Hawthorne's The Scarlet Letter might disabuse you of that notion).
In any event, Europe recently enacted a set of rules under the concept of "the right to be forgotten." These rules require search engines, such as Google Search, to offer the European public a process for requesting the removal of specific search results that appear after specific search terms are entered. Specifically, stigmatizing information (that's not also in the public interest) must not appear when a search is conducted for the name of a person who has requested the removal.
For example, let's say a European man is photographed by the local newspaper while playing an Australian didgeridoo in the park. He's young, bearded and has long hair. Years later, he sells the instrument, shaves his beard, cuts his hair and becomes a respected financial adviser. But when prospective clients search Google for his name, they find the old newspaper photo. The man can request that Google remove the link to the newspaper when people search his name, and Google is required by law to comply. (If that sounds far-fetched, note that I'm describing an actual case under the right-to-be-forgotten law.)
Google has received hundreds of thousands of requests for right-to-be-forgotten removals, and it has granted nearly half of them.
The Russian parliament has approved an even stronger right-to-be-forgotten measure which, if signed by President Vladimir Putin, will be become law next year.
And last week a group called Consumer Watchdog asked the Federal Trade Commission to enact right-to-be-forgotten rules in the U.S.
A search engine exists to faithfully index and enable the discovery of legal content posted online. But right-to-be-forgotten rules have the effect of selectively rendering search engines less accurate.
Internet content can stigmatize people, and that's a problem. The right-to-be-forgotten solution requires Google to fix that problem by sabotaging its search engine and offering an inferior product.
Criminals can use various methods to communicate with one another so police can't listen. For example, they can whisper in each other's ears, talk in code or use burner phones. And, of course, they can use end-to-end encryption for texting or emails.
While bad guys could and do use encrypted communication to jeopardize national security and to commit crimes, encryption is still very useful as a tool to protect national security and to prevent crime. Without encryption, it's much easier for hackers, including state-sponsored perpetrators of industrial espionage, to steal military, business and other secrets and use the information against U.S. organizations. Encryption also helps prevent a long list of crimes, including identity theft, extortion, blackmail and other kinds of fraud.
But the FBI and the U.S. Department of Justice recently announced that Google (among other companies) is a major threat to national security and law enforcement because it provides end-to-end encryption.
The solution, according to law enforcement organizations, is for Google and other companies to create back doors that would enable authorities to eavesdrop on encrypted communication.
The trouble is that any back door created for law enforcement could also be used by terrorists, criminals and hostile foreign governments for hacking, spying and stealing. Security experts know this; the FBI and the DOJ apparently don't.
Yes, there is crime. And the solution proposed by federal law enforcement authorities is for Google to fight crime by making its products unsecure. That solution prevents Google from meeting the public's demand for secure communication.
An unrelated situation involves Google's social mapping app, Waze. The free app is like Google Maps, except that it enables people to communicate with each other in specific ways. Drivers can use Waze to alert one another about things like road hazards and traffic jams. They can also use it to identify the location of police cars.
Police organizations recently called on Google to disable the ability for Waze to note the location of police cruisers, saying it enables the stalking of police officers. They want Google to prevent people from communicating with each other about police cars they see while driving around (something the police and the government themselves cannot prevent because of the First Amendment to the U.S. Constitution).
In yet another case, Google is being called upon to address the problem of youth violence.
Google's personalized advertising system on search and YouTube grabs "signals" to decide which ads to display. These signals come from both the content displayed and also the search history and other personalized information of the user.
Because rap music often talks about guns and violence, ads for gun-related companies and organizations are more likely to be displayed in the context of rap content and videos.
The founder of an organization called the Hip-Hop Chess Federation demanded that Google intervene in its own ad-serving algorithm to remove ads for the U.S. Concealed Carry Association (USCCA) and other gun-related companies on rap-related content.
In New York City last year, 20 people were killed by vehicles that were making legal left turns, according to this WNYC article published two months ago.
Because of the article, the New York City Council this month wrote a letter to Google asking the company to make two changes to Google Maps.
One of the council's requests is for a "stay on truck routes" feature for truck drivers. The other is for functionality that would give Google Maps users the ability to request turn-by-turn directions with fewer left turns.
Car accidents happen. It's a real problem, and it's tragic for victims and their families.
Google is already doing as much as any organization to address the problem by inventing a self-driving car that's way safer than human-driven cars.
But the New York City Council now wants Google to change its Maps app in order to solve society's car accident problem.
Here's the problem. Some or all of these ideas might be worth doing. They might make the world a better place. They might not.
But taken together, they have the overall effect of degrading the quality of Google's services, watering down Google's mission and turning Google into a company that intervenes in the way people communicate with each other.
It's social engineering.
The world has problems, but is breaking Google really the best way to fix society?
Google's mission is to organize the world's information, not rearrange the world's information to solve the society's problems.