First thing first - what is AI?
Artificial intelligence
Artificial Intelligence (aka AI) leverages computers and machines to mimic the problem-solving and decision-making capabilities of the human mind.
You can't really escape it. Most of us get in touch with it on a daily base. From face recognition to unlock your phone to using Siri or Alexa to helping get rid of social spam on the internet.
How would you describe an social spam?
Wikipedia defines social spam as – “Unwanted spam content appearing on social networking services, social bookmarking sites, and any website with user-generated content (comments, chat, etc.).” In some cases, toxic comments online have even resulted in real life violence, from religious nationalism in Myanmar to neo-Nazi propaganda in the U.S.
EXAMPLES OF AN SOCIAL SPAM:
- BULK MESSAGES
- PROFANITY
- INSULTS
- HATE SPEECH
- MALICOIUS LINKS
- FRAUDULENT REVIEWS
- EXPLICIT PICTURES
- THREATENING POSTS
How to fight against unwanted spam content?
PEOPLE
Traditionally, it was human labour, which was heavily unhealthy. The occupation was notoriously terrible, psychologically injuring workers who have to comb through the depths of depravity, from child porn to beheadings. Burnout happens quickly, workers cite symptoms similar to post-traumatic stress disorder.
In 2019, it was reported that Facebook moderators are at risk of suffering from PTSD as a result of repeated exposure to such distressing content.
Outsourcing the "dirty" work to machine learning can help manage the rising volumes of harmful content, while limiting human exposure to it. Indeed, many tech giants have been incorporating algorithms into their content moderation for years.
Let's take a look:
According to the Mountain View firm’s teams, these solutions can block up to 99% of unwanted content. About 40 billion spam pages are discovered every day among the billions of sites or mails explored and indexed. Let’s take a look at the techniques used by Google to eradicate spam and malicious content.
For instance, it introduced a keyword filter that automatically removed some of the offensive words that appeared on the feed. The filter worked on the mechanism of removing words that were “often reported as inappropriate.” Moreover, it also targeted custom keywords set by the users. The system was also trained to work on at least two million comments that were categorized in various segments, including racism and bullying.
The same system also assists in spam detection by using data assets and human input. It identifies the fake accounts and works to clean the comments posted by them on various posts and videos.
it relies on AI to automatically detect fake accounts based on the number of accounts per device. The mechanism is simple – Facebook’s AI labels an account as a bot when it signs up and sends more than 100 friend requests in a span of a minute.
Facebook also expanded a fact-checking program to include images and videos. The tools used for this purpose used AI of AdVerif.ai. This helps the platform to find flagged images and then go for a reverse image lookup and proceed to see alterations.
YOUTUBE
This step can filter the toxic comments and curb online harassment that ruins conversations. The AI quickly gathers information from the labels, and the moderators can then take action. Ergo, they can filter out comments that are declared toxic by the algorithm.
For now, automation does play a role in certain situations, but the users on the platform are helping to train the AI. The platform gathers data on how often any account is muted, blocked, reported, retweeted, liked, or replied to. This is how the AI mechanism can recognize an account, which has been already blocked by a large number of people and flag it for the moderators to take action.
The AI is also able to distinguish between both negative and positive interactions. It helps the platform curate experience for the users.
BUMBLE
First debuted in 2019, Private Detector recognises and blurs lewd images with 98% accuracy, that are sent through the Bumble app, giving the user on the receiving end the choice of whether to open, block or report the image. A 2017 survey from YouGov found more than 50% of millennial women had received an explicit image electronically, and more than 75% of these recipients had not sought them out. (Interestingly, less than 25 percent of men surveyed admitted to having sent such photos.) The feature also prevents explicit images from being uploaded to user profiles.
However.....
The use of AI, particularly by big tech platforms and other companies, comes with limitations and problems of its own. The report outlines significant concerns that AI tools can be inaccurate, biased, and discriminatory by design and incentivize relying on increasingly invasive forms of commercial surveillance.
“Our report emphasizes that nobody should treat AI as the solution to the spread of harmful online content,” said Samuel Levine, Director of the FTC’s Bureau of Consumer Protection. “Combatting online harm requires a broad societal effort, not an overly optimistic belief that new technology—which can be both helpful and dangerous—will take these problems off our hands.”
In legislation enacted in 2021, Congress directed the Commission to examine ways that AI “may be used to identify, remove, or take any other appropriate action necessary to address” several specified “online harms.” The harms that are of particular concern to Congress include online fraud, impersonation scams, fake reviews and accounts, bots, media manipulation, illegal drug sales and other illegal activities, sexual exploitation, hate crimes, online harassment and cyberstalking, and misinformation campaigns aimed at influencing elections.
Overall...
AI can help us a lot, especially doing the work that is exhausting for humans. However, we cannot forget that it has its own limitations. If AI is to stand a chance of replacing manual effort on a large scale, it is crucial that we give our models the full picture.
Credits:
https://techcrunch.com/2022/10/24/bumble-open-sourced-its-ai-that-detects-unsolicited-nudes/?guccounter=1
https://www.washingtonpost.com/business/2019/04/25/tired-being-deluged-with-unsolicited-nudes-bumble-says-private-detector-will-protect-you/
https://techcrunch.com/2016/05/31/terminating-abuse/
https://www.actuia.com/english/how-does-google-seek-to-fight-unwanted-content-with-artificial-intelligence/
https://www.scientificamerican.com/article/can-ai-identify-toxic-online-content/
https://www.linkedin.com/pulse/how-top-social-platforms-using-ai-fight-spam-gauri-bapat/
https://www.ftc.gov/news-events/news/press-releases/2022/06/ftc-report-warns-about-using-artificial-intelligence-combat-online-problems
https://www.ibm.com/cloud/learn/what-is-artificial-intelligence
Credits:
Created with an image by Prostock-studio - "Psychotherapy, treatment of depression concept. Stressed middle eastern guy having session at psychologist's office"