Online predators create and share the illegal material, which is increasingly cloaked by technology. Tech companies, the government and the
If last year's Tumblr Purge had actually been effective, it wouldn't have effectively gone after all the CSAM shit instead of going after artists, photographers, writers, models, actors, and their fans.
However, Verizon much like any other company invested in technology took a shotgun approach to a scalpel problem. The result is the "porn" problem got mismanaged while the investigators saw the CSAM problem get worse.
The problem is digital bias, and it's the same issue that the LGBT community made a 31 minute video on YouTube exposing how YouTube and Google screwed the pooch when it came to machine learning by hiring low-wage workers in ultra-conservative countries to make decisions on what content could be monetized.
One person's bundle of twigs was another person's terrible insult. A reminder to all you ML/AI geeks: Robots do not understand insults, sarcasm, or even bias very well. They will take everything literally.
What's more, and this is for all the conservatives out there who think "liberals are censoring conservative free speech", you would be incorrect. If a group of folks like the LGBT community, which for the most part outside of the "Log Cabin" group of conservatives is being demonetized by robots, what do you think is doing the same thing with all you stuff? The same thing.
The robots aren't "liberal" they are LITERAL.
So every vulgar Obama joke pretty much gets processed like any bad Trump meme...or so we thought.
In fact the robots are Racist AF! They learned their hatred like anybody else: a consistent diet of being exposed to bias shit.
Microsoft's Tay.ai was ruined because what started as a joke by some 4chan pranksters turned into a nightmare.
One computer scientist, Joy Buolamwini noted that because how Machine Learning has trouble with recognizing faces of people with darker complexions, the machines don't recognize them, which for her, a person of color, was quite disheartening.
Algorithmic Data Bias doesn't just discrimination against race, gender, sexual activity, but also crime statistics, justice, financial lending, disability, healthcare, immigration, even hiring decisions.
Ever get turned down for a job to find out somebody not as good at the job as you gets hired or promoted but later it's discovered the person they hired was corrupt? Imagine AI or ML doing this constantly.
The bots make the decisions, but the parameters for how those decisions are made are still set and modified by humans, many with bias opinions either by a lack of understanding or education. In some places, state-run propaganda as been applied not to apply data to a specific democracy, but to be applied universally beyond borders. We saw this with how China tried to influence Google, how Russia did it with elections in the Ukraine and the United States, how Verizon and Yahoo did it when they ran Tumblr.
By now, we should know better than to apply Algorithmic Bias in data. What may seem like a good idea for one group to squelch out their rivals and enemies overall excludes, isolates, or ostracizes ther few supporters and allies that exist in the same area.
The one effective means to defeat Algorthmic Bias is to infiltrate it.