Human content moderators are coming back to work on YouTube because the AI systems could not make nuanced decisions as a human mind does

Elon Musk, the CEO of SpaceX and Tesla often says that Artificial Intelligence would soon take over the human races and the world will become a place controlled by machines and technology, pushing humans way back. However, other tech magnates keep refuting these claims and YouTube has recently proved that nothing can beat human intelligence, and human beings will always remain above everything.

When the coronavirus pandemic hit the world, many things happened that had never occurred before. Due to global lockdowns and social distancing measures, millions of people started the new trend of work from home. YouTube also had to reduce its staff and workload of most of its in-office employees, which included a large number of content moderators. The job of content moderation was then assigned to automated systems, AI technologies that were able to detect harmful content in videos on YouTube and then remove them promptly. These systems gave a broader aspect of content-moderation in the place of around 10,000 human moderators, and initially, it was considered as a wonderful approach too, besides being the need of the hour.

However, it turned out that the AI systems removed around 11 million videos from April to June 2020, which was an unusually high number of videos supposedly containing harmful content. When YouTube’s teams looked into the matter, it was found out that YouTube’s AI systems made some error and removed a lot of videos that actually did not contain any harmful content, neither they violated any of YouTube’s guidelines. YouTube had to review these videos and recover 160,000 ‘innocent’ videos on the platform. YouTube hardly reverses its ruling and moderation decisions on less than 25% of appeals, but this time, around 50% of those appeals were reversed because of the blunder of YouTube’s AI systems.

YouTube’s chief product officer Neal Mohan has recently told Financial Times that the world's biggest video search engine is now re-assigning the content moderation task to human beings again because of this incident.

Neal Mohan says that this was kind of expected because YouTube knew that Artificial Intelligence is not equal to Human Intelligence, and machines cannot make precise decisions like humans do. Therefore, now that the world is trying to get back on its feet and the lockdowns have lifted in many regions, YouTube is also bringing its human content moderators back to their offices so that they can make more nuanced decisions. However, since the COVID-19 is still there, it is not known when the decisions will reverse and when something new happens. So, Mashable is keeping an eye on any further news about this issue from YouTube.


Photo: Toby Melville / reuters

Read next: Mozilla’s New Browser Extension Might Force YouTube to Improve Its Recommendation Engine

No comments:

Post a Comment