News

AI, fake news, and future concerns

“What no one knows is how misinformation will spread during the election season this year,” says Jency Jacob, Managing Editor of BoomLive. Investigating bogus news that circulates on the internet and social media is what Jacob and his crew do. This year, we had a discussion on the function of Big Tech businesses. Voting for the next head of state will take place across half of the world, including India. This year has been unlike any other in the history of the world’s democracies, by any measure. Jacob argues that although he and his group have long been fighting false information, the past year has seen a growth in artificial intelligence (AI), which has complicated matters and left most people unsure of what to anticipate.

Read More: AOX News

“Apart from the typical narrative disinformation that transcends party lines, the biggest obstacle this election season will be how they leverage new technology like artificial intelligence. Artificial intelligence (AI) technologies produced a large amount of copied audio during the most recent state elections, which was used to trick voters. The technology is becoming increasingly complex and difficult to identify with every day that goes by, he claims.

“We had cheap fakes in India, but not deep fakes.” Jacob uses an audio video clip that went viral in the latter part of last year to support his claims. It featured Amitabh Bachchan asking a participant in the well-known show Kaun Banega Crorepati a question in a voiceover. The response portrayed former MP chief minister Shivraj Singh Chouhan as a criminal, which made the information defamatory. The footage was posted by Ritu Choudhary, a top Congress politician, albeit it remained unknown who developed it and made it public. Furthermore, the harm had already been done when it was discovered to be a phony. It had already been viewed by a couple million individuals, many of whom had fell for it.

According to Jacob, in order to detect deep fakes, his team closely collaborates with the disinformation team at Meta, which is home to Facebook, WhatsApp, and Instagram. He’s not sure, though, how much of the misinformation that the team flags is subsequently removed. “We are given to understand those pieces of content are suppressed,” Jacob explains warily.

However, based on their lack of cooperation, he feels that Google is the most annoying company to work with. The fact that their algorithms are unclear just makes matters worse. Because Google has almost complete control over the search engine business, this is very noteworthy. It is also the owner of YouTube, which users utilize to obtain video material. He claims that “they frequently don’t take down misleading content because they are worried about who they might annoy.”

However, Google rejects these claims. A representative said, “Not at all—the baseline approach to our recommendation system is built on principles that have stood the test of time,” when questioned about neglecting content filtering. Additionally, our systems change as new threats and technology arise.

He was just as reluctant to admit that, with their crew now smaller, content moderation had suffered. “Our systems are updated often, and since we operate on such a large scale, moderation always combines automatic and manual review techniques. Therefore, I wouldn’t say that labor cutbacks and the effect on accountability efforts are directly related.

Rather, he argues that YouTube has consistently enforced Community Guidelines using a mix of human and machine learning technology, with over 20,000 reviewers working remotely for Google. “In our systems, reviewers verify if material has breached policy boundaries while AI classifiers assist in the large-scale detection of potentially offensive content. Furthermore, AI is constantly improving our content filtering systems’ accuracy and speed.

What then is a reasonable expectation about the dissemination of false information? “Takedown notices could increase,” stated a senior Google employee who has since moved on and did not want to be identified. There will be more backdoor talks and pressure to suppress unfavorable news articles. Even if Google doesn’t cooperate, there will undoubtedly be pressure. Politicians refer to democracy while they are not in office. They look up to Chinese autocrats while they are in power.