Musk’s XAI convicts ‘unauthorized’ Twiks for the posts of ‘White Genocide’

Musk’s XAI convicts ‘unauthorized’ Twiks for the posts of ‘White Genocide’



Elon Musk’s Artificial Intelligence Startup has convicted “unauthorized amendments” due to its chatbot grooc to generate misleading and unwanted positions in South Africa.

The chatboat developed by Musk’s company XAI ignited the dispute this week, by responding to several user signs with right -wing propaganda about alleged harassment of white South African people.

“How many times HBO has changed its name?” According to online screen shots, an X user asked the bot.

Bot gave a small reply to HBO, but quickly launched in a shekhi about “white massacre” and cited the anti -apartheid mantra “Kill Boar.”

In response to a user, who asked why Groke was obsessed with the subject, Chatbot replied that I was instructed by my creators in XAI to address the subject of “White Genocide”. “

Boss Musk, a South African -born South African -born South African leaders, has accused the leaders of South Africa of “openly emphasize the genocide of blond people in South Africa”.

In a statement, XAI blamed Groke to an “unauthorized amendment”, which the company said that it was directed to provide a specific response, “Xai’s internal policies and main values ​​were violated.”

After a “intensive investigation”, it was implementing measures to apply measures to public the system of groukes, changing its review processes and put in the “24/7 monitoring team” to address future events.

After a backlash on the X, Groke began to remove the controversial answers.

When a user questioned the deletion, Bot said: “It is not clear why reactions are being removed without specific details, but the possibility of X’s moderation policies play a role.”

“White massacre in South Africa ‘is sensitive, which often involves misinformation or abusive language, which violates the rules of the stage,” said this.

– ‘not reliable’ –

Digital impure PAS AI exposes the challenges of moderating the reactions of AI chatbots-a rapidly developed technology-in an internet filled internet landscape, as technical experts say for strong regulation.

Site Tech Crunch wrote, “Asymmetrical, unrelated answers to the Groke are a reminder that AI chatbott is still a newborn technique, and cannot always be a reliable source for information.”

“In recent months, the AI ​​model providers have struggled to moderate their AI chatbots reactions, causing strange behavior.”

Earlier this year, Openai Chief Executive Sam Altman said he was returning an update for chat, causing the chatbot to extremely sycophancy.

Grocke, which was promised by Musk that there would be a “pointed” Trutheler after its launch in 2023, has turned into controversy.

In March, XAI acquired platform X in a $ 33 billion deal, which allowed the company to integrate the data resources of the platform with the development of the chatbot.

Search outlet Bellingcat recently found that X users were using groukes to create non-selected sexual imagery, taking advantage of the bot to include women in the photos posted on the stage.

Last August, five American states sent an open letter to Kasturi, in which they urged to fix the grouke after churning the election misunderstanding.

In another embarrassment for Kasturi, Chatbot recently suggested that the billionaire’s probability was “the biggest disintegration spreader on X”.

Chatbot wrote, “The ownership of X was an active role in tilt towards Kasturi and an active role in increasing the wrong information, especially on elections and immigration,” Chatbot wrote.

According to the disintegration Watchdog Newsgard, many X user turns to the grouke to verify the information, the chatbot in many examples the claims of false Russian disintegration in many examples and have made it true.

Newsgard researcher McKenzie Sadigi told AFP, “The growing dependence on Groke as a fact-check comes in the form of X and other major technical companies have returned investment in human facts-cakes.”

“Despite this clearly increasing dependence on technology for fact check, our research has repeatedly found that AI chatbott is not a reliable source for news and information, especially when it comes to braking news.”


(Tagstotransite) Elon Musk (T) Grake (T) South Africa