Facebook founder Mark Zuckerberg said Wednesday that identifying hate speech on the platform through artificial intelligence is a far more difficult task than doing the same for a nipple.
The tech executive’s candid comments came during a conference call with analysts after his company revealed its first-quarter earnings report, according to Business Insider.
Results in the first portion of the year were fairly positive, as Facebook posted revenue of $12 billion, a substantial jump over the same quarter last year, and its number of users increased by 3.5 percent (49 million).
Somewhat like the updated statistics, Zuckerberg’s remarks offer a glimpse into the corporation’s well-being and mindset.
“It’s easier to build an AI system to detect a nipple than what is hate speech,” he said, according to Business Insider.
The wunderkind’s line of thinking seems to show that he has reservations about the company’s ability to fairly police content on the platform since determining hate speech is subjective and thus inherently difficult. Identifying offensive subject matter and communications is prone to the differing personalities of the individual or groups of moderators, leaving AI to deal with the purported problem.
But just as it is tough for humans to impartially monitor the billions of pieces of content on the massive social media network, so too is it hard to create an AI system that doesn’t reflect any innate biases of their creator.
Nevertheless, Zuckerberg said or implied multiple times during his highly anticipated congressional hearings that AI is really the only choice for the ultimate goal of making the platform as clean as possible.
Since those hearings, which followed a number of recent events and ostensibly new revelations, Facebook has tried to be more transparent about how it operates. For example, it revealed its once hush censorship rules Tuesday, detailing how it cracks down on content like “graphic violence,” “cruel and insensitive” content and behavior like bullying, “adult nudity and sexual activity,” and “hate speech.”
Each category provides criteria that can be interpreted and therefore enforced in many various ways. When an AI system is the entity doing the deciphering and subsequent implementation, the same is possible, apparently corroborating Zuckerberg’s claim that nipple spotting is far less complex than ambiguously defined hate speech.
But even drawing the line for nipples isn’t easy.
“We restrict the display of nudity or sexual activity because some people in our community may be sensitive to this type of content,” the company’s community standards read. “Our nudity policies have become more nuanced over time.
“For example, while we restrict some images of female breasts that include the nipple, we allow other images, including those depicting acts of protest, women actively engaged in breast-feeding, and photos of post-mastectomy scarring.”
As calls to do more by portions of the population increase over the years, Facebook has been constantly accused of inappropriate censorship — including that of content in some way displaying a female body part.
The company on separate occasions in recent years rejected images of a woman’s bare back, blocked an instructional breast examination video that showed animated boobs, and restricted the social media capabilities of a photographer who records mothers giving birth.
A version of this article previously appeared on The Daily Caller News Foundation website.
Truth and Accuracy
We are committed to truth and accuracy in all of our journalism. Read our editorial standards.