Monday, October 22, 2007

...Google's problems in Brazil with inapproriate content...
The Wall Street Journal recently published a front page article highlighting Google’s problems in Brazil with inappropriate content on its social networking. Orkut, Google’s version of social networking sites similar to and NewsCorp’s, in an inexplicable scenario best explained by Malcolm Gladwell’s book The Tipping Point; has become wildly popular with Brazilians. Social networking sites allow users to establish online communities (topics) and upload content to both the community areas on the servers and to personal web sites (which may also have community affinities).

Anyway, leaving the intricacies of the platform aside, the communities created range from those dedicated to praising moms to the darker side of human nature: racism, pedophilia, animal cruelty and so forth. The WSJ does its usual excellent job of explaining how Google got in hot water, so I won’t rehash that. What I want to point out is the root problem and a proposed solution.
The root problem is...
The root problem is that these social networking sites (like any internet site) allow users with access to upload anything. In the case of Google’s Orkut, all content uploaded ends up on Google’s servers in the USA, where free speech is an inalienable right. Not so in Brazil. So the real questions are:

Do content hosts have a legal right or obligation to edit content?
What laws apply?

Let’s deal with the legal right to edit content first. Google, and other content hosts, have what is called Terms of Service (TOS). In these, they specify the types of content that is not approved for uploading to their servers. If a user violates these terms, then Google has the right to delete the content and remove the user account under the TOS because the user agreed to the TOS in order to receive an Orkut account.

Next, does Google have an obligation to remove content that violates its Orkut TOS? Well, according to the TOS, they do. However, they also state they are not liable for content that violates the TOS. That’s a topic for another day, so let’s assume for the moment they have an obligation but the timeliness of their obligatory response is fuzzy.

If a Brazilian citizen in Brazil uploads pro-Nazi comments or racial slurs...
Now we get to the meat of today’s topic. What laws apply? If a Brazilian citizen in Brazil uploads pro-Nazi comments or racial slurs and the uploaded content is stored on, and served from, computers in the US, whose laws apply if the content is subsequently served to another Brazilian citizen in Brazil? What if the same content is uploaded by a US citizen standing on US soil? By a Brazilian on US soil? By a US citizen in Brazil? .

Logically, if the content is illegal in Brazil and was uploaded by a Brazilian who was actually in Brazil at the time, the Brazilian citizen seems to be the culprit and the Brazilian government within its rights to prosecute that citizen. Yet the content is stored and served from computers in the US, where such content may be perfectly legal. Because the content can be served to citizens in their homes and offices in most countries of the world, which have varying laws and standards, the permutations become seemingly infinite.

Unfortunately, when you offer a free service...
The solution is crafting adequate Terms of Service and enforcing them. If Google (and others) craft TOS agreements that reflect the laws of the country of origins where they allow users to sign up, then Google and the others should enforce those agreements by screening content before it is available to the public. Unfortunately, when you offer a free service and users upload millions of items daily, enforcement comes at a high cost. Human review of the millions of images alone uploaded daily would be very expensive. When you throw in text posts, the number of editors needed would start to look like the cadre of TSA screeners.

You also have to consider there may be legal content in one country that that country’s TOS allows that is illegal elsewhere and users signing up in that country have different content standards. So Google may want to display that content to users in one country but lock it to users in another.

Some would say police nothing and allow everything...
So what is the solution? Some would say police nothing and allow everything, however, I think we can safely assume that society is not going to accept that alternative. So the solution has to lie with technology, of course. Specifically, the solution is applying image recognition software to screen out photos not meeting the TOS standards and using context sensitive text screening software to catch the words of extremists and pedophiles who violate local law or the TOS. Since the sophistication of existing software is not up to the task, this is not an easy solution. However, it is probably easier than getting all the countries of the world to agree to uniform laws or to a global internet authority. Google are you up to the technical challenge? BTW, the Homeland Security folks would love the software, too, so there may be a big paycheck in it for Google!