Tesla CEO Elon Musk claims that his aggressive takeover offer for Twitter is all about free expression. On the matter, Tesla CEO Elon Musk has posted many surveys to his nearly 80 million Twitter followers, and he stated this week that he thinks the social media firm is a town square that can promote world peace, provided that there is little policing. He wants to make the computer code that decides what users view on the Twitter public in order to guarantee that the site is fair if he is successful in his $43 billion acquisition.
The algorithm decides how many millions of people view tweets based on how often they are re-tweeted. So he said, he wants to make the algorithm that determines whether a tweet is promoted or demoted “open source,” which means that everyone can see and improve on it. “Behind the scenes manipulation” would be prevented, he added. He added in a live webcast from the TED conference on Thursday that whether tweets are “emphasized or de-emphasized, that action should stay obvious.”
According to specialists who study recommendation systems, Musk’s idea likely constitutes a huge simplification of how it would function to make such data public. Social media firms’ recommendation engines’ underlying software has become so large and complicated that analyzing it would need access to an enormous amount of data that most individuals don’t even have access to. Algorithms on sites like Twitter, Facebook, and other social media sites analyze billions of bits of information to determine a post’s importance.
Recommendation engines’ algorithms aren’t one thing, says Nick Seaver, an anthropology assistant professor at Tufts who studies the mechanisms behind them. The processes are so complicated, he claims, that even computer firms struggle to understand why a user was shown a certain message over another by their software.
It’s not only outsiders that are curious about Twitter’s algorithm, but he also said. Musk, Twitter, and Facebook all declined to comment when contacted for comment.
Conservatives, in particular, have long been fascinated by the opaqueness of algorithms that feed what users see on social media sites. They have claimed without proof that the platforms are skewed against them. Some on the right have taken Musk’s efforts to seize control of Twitter this week with glee, believing that he may be attempting to lessen the policing of false information and reinstate former US President Donald Trump to the site.
Efforts have been made to regulate algorithms so that they are more transparent. New legislation in the US Congress last year focused specifically on the software programs that determine what individuals see on social media sites, including Facebook, Twitter, and Instagram. AI and algorithm regulation are likewise developing over the world.
A complicated algorithm, which is basically a set of principles like a mathematical equation that help comb through information and identify what is most relevant to an individual user, wasn’t necessary when social networks like Facebook and Twitter were first getting started.
Companies started creating algorithms that could learn which items users were more likely to click on and then filter their feeds appropriately when hundreds of millions of people joined and posted billions of pieces of information.
It’s now common practice for corporations such as Facebook to utilize some kind of algorithm to choose what to display to its consumers. So, not everyone on a social network sees the same thing, which is a problem. A basketball lover may be inundated with postings on the NBA playoffs, while those who are outdoorsy may receive recommendations for the greatest summer camping places.
Because of concerns that their ranking methods aid in the propagation of disinformation, social media platforms Facebook and Twitter have lately been under pressure to restore chronological feeds as the default setting.
Machine learning is a fancy name for pattern recognition, which is what most corporations term the complicated arithmetic they utilize. Computers cannot identify whether a tweet is humorous, entertaining, or useful just on the content of the tweet. To some extent, though, it is possible to make some educated guesses about which tweets are most likely to get attention based on an analysis of millions of tweets and other criteria such as who liked, shared, and retweeted them.
Michael Ekstrand, an assistant professor of computer science at Boise State who studies recommendation engines, says that even if Twitter made its formula public, including the math it uses to “train” its machine learning algorithms, an outsider looking at it wouldn’t be able to make meaningful conclusions from it.”
Additionally, an outsider would require access to the data used to “train” those algorithms – the millions of tweets that have been seen, liked, or shared.
Experts say that disclosing this information would give rise to grave privacy issues. When it comes to recommendation algorithms, “the algorithm is only a starting point,” says Robin Burke, an expert in the field. According to him, “the remainder of the iceberg is all of this data that Twitter possesses,” with the majority of it being restricted from public disclosure.
According to the experts, Musk should be familiar with the intricacies of algorithms. He is the CEO of Tesla, a business that is developing self-driving cars using machine learning algorithms. Tesla is creating its own supercomputer and bespoke semiconductors to handle all the data generated by this project.
According to the experts, there are other, more practical methods to improve openness, some of which Twitter currently does. It has been suggested by some opponents that social media corporations should simplify their algorithms in order to eliminate apparent biases against certain categories of individuals. Others have advocated for internal audits conducted by non-profit organizations.
Twitter has an in-house research team called Machine Learning Ethics, Transparency, and Accountability (MLEA) that investigates algorithmic biases and provides transparency and accountability. Among other things, it has released a study on whether the algorithms used to automatically trim profile images were biassed in any way.
Twitter has been suggested as a platform for several algorithms. It is possible for Twitter to pick commercial partners who would have access to the data and build algorithms customized to certain audiences. ”
Co-founder of the Every writer’s group Nathan Baschez stated in a post on Friday that Twitter should enable outsiders to design their own algorithms that are targeted toward certain interests. In response to a tweet from Baschez, Twitter co-founder Jack Dorsey said that he had pitched the notion when serving as the company’s chief executive before stepping down in December.
In November 2020, Dorsey made the suggestion in a Senate hearing. We need more options when it comes to how algorithms are used, he said. “Algorithms are responsible for showing us what we see or what we don’t see,” he said. One of Bluesky’s goals is to offer individuals greater control over how their feeds are organized. Twitter is also supporting Bluesky.
On condition of anonymity, a former Twitter employee claimed the business had contemplated creating an “algorithm marketplace” where users could pick and choose how they want to see their feeds reorganized. However, the individual added attempts to provide greater transparency have been difficult because of the way Twitter’s algorithms are intertwined with other areas of the system. Opening it open might expose confidential information and attract misuse, according to the source.
According to Burke, the notion has appeal but would need a reorganization of Twitter’s operations and data flow. There are too many social media giants for that to happen, and it only goes to show how monopolistic they are.
Additionally, releasing Twitter’s software to the public may lead to undesirable consequences. Experts argue that those who are trying to manipulate the system by disseminating misinformation to influence an election might use this knowledge to do so.
No matter how much success Musk has, he will still face questions about whether the code he provided publicly is complete and if anything was left out. According to the millionaire, he would be relying on Twitter users to believe that there was no shady activity taking place behind the scenes. Skepticism and cynicism can’t be completely eliminated.” According to Ekstrand, “There will always be those who don’t believe what’s being stated.”
Source: The Washington Post