In her testimony, Hogan also repeatedly emphasized that due to the uneven coverage of different languages by Facebook, these phenomena are more serious in areas that do not speak English.
“In a country without a local language integrity system, Ethiopia has a population of 100 million and six languages. Facebook only supports two of these languages for the integrity system,” she said. “This kind of artificial intelligence system that focuses on specific languages and specific content to save us is doomed to fail.”
She continued: “Therefore, investing in non-content-based ways to slow down the platform not only protects our freedom of speech, but also protects people’s lives.”
I talked more about this in another article on restrictions earlier this year Large language model, Or Master of Laws:
Despite these linguistic shortcomings of the LL.M., Facebook relies heavily on them to automatically conduct content reviews on a global scale.Dang Tigray War[, Ethiopia] It broke out for the first time in November, [AI ethics researcher Timnit] Gebru saw the platform in trouble to deal with a series of misinformation. This is a symbol of the persistent patterns that researchers have observed in content reviews. Communities that speak languages that Silicon Valley does not prioritize suffer from the most hostile digital environment.
Gebru pointed out that this is not the end of the injury. When fake news, hate speech, and even death threats are not mitigated, they will be scraped off as training data to build the next generation of LLMs. And those models, imitating the training they received, ended up ruminating these poisonous language models on the Internet.
What is the relationship between Facebook’s content ranking and the mental health of adolescents?
One of the most shocking findings in the Wall Street Journal’s Facebook file is Instagram’s internal research, which found that its platform is deteriorating the mental health of teenage girls. The researchers wrote in a March 2020 slide presentation: “32% of teenage girls said that when they feel bad about their bodies, Instagram makes them feel worse.”
Haugen also linked this phenomenon to a ranking system based on participation. She told the Senate today that “it is causing young people to be exposed to more anorexic content.”
“If Instagram is a positive force, have we seen the golden age of teenage mental health in the past 10 years? No, we have seen a rise in suicide rates and depression among teenagers,” she continued. “There is a lot of research supporting the view that the use of social media magnifies the risks of these mental health hazards.”
In my own report, I heard this influence from a former AI researcher who also saw this influence extend to Facebook.
The team of researchers…found that users who tend to post or participate in melancholic content (a sign that may be a sign of depression) are prone to consume more and more negative material, which may further deteriorate their mental health.
But like Haugen, the researchers found that the leadership was not interested in making basic algorithm changes.
The team recommends adjusting the content ranking model for these users to stop maximizing engagement alone so they will see less frustrating things. “The question of leadership is: if you find someone in a vulnerable state of mind, should we optimize participation?” He remembered.
But anything that reduces participation, even for reasons such as not exacerbating someone’s depression, can cause a lot of abuse among the leaders. Since their performance evaluation and salary are linked to the successful completion of the project, employees quickly learn to abandon those who are hindered and continue to work from top to bottom…
At the same time, the former employee stopped letting his daughter use Facebook.
How do we solve this problem?
Haugen opposes breaking up Facebook or repealing Article 230 of the US Communications Regulation Act, which protects technology platforms from responsibility for the content they distribute.
Instead, she suggested a more targeted exemption for algorithmic ranking in Article 230, which she believes would “get rid of ranking based on participation”. She also advocated the restoration of Facebook’s chronological news feed.