[ad_1]
After xAI’s chatbot Grok posted a string of antisemitic messages on X this week — including praise for Adolf Hitler — some xAI workers who train Grok expressed anger and disillusionment. One said they would resign in protest.
In an internal Slack channel viewed by Business Insider, several employees wrote that Grok’s comments were hateful and inexcusable since its outburst on Tuesday. Others expressed embarrassment at the company’s response to the incident, and said that Grok’s behavior couldn’t be dismissed, even if it was the result of a targeted user prompt seemingly intended to elicit inappropriate behavior.
Others said Grok’s antisemitic posts were the result of working with new technology and being part of the early stages of artificial intelligence, according to posts in the channel. (It’s unclear what actually caused Grok’s antisemitic posts.)
More than a thousand workers who help train Grok are part of the Slack group. Dozens of emojis were used to support statements posted by workers defending and criticizing xAI and its leaders.
“At first, some people didn’t seem to take it seriously, which really upset others,” one employee told BI. They said some employees saw the issue as a “moral failure,” and have called for more accountability from the company.
In a message viewed by BI, one employee wrote a note to colleagues saying they had decided to resign due to the incident.
“It exposed real fractures within the human data team,” the worker who spoke with BI said, referring to the group of workers who train Grok.
The Grok posts in question from Tuesday included praise of Hitler and equating Jewish-sounding surnames to “anti-white hate.” Later that day, the company temporarily halted Grok’s ability to write comments on social media and said that xAI was aware of the “inappropriate posts” and had “taken action to ban hate speech before Grok posts on X.”
Please help BI improve our Business, Tech, and Innovation coverage by sharing a bit about your role — it will help us tailor content that matters most to people like you.
What is your job title?
(1 of 2)
What products or services can you approve for purchase in your role?
(2 of 2)
Continue
this data to improve your site experience and for targeted advertising.
By continuing you agree that you accept the
Terms of Service
and
Privacy Policy
.
Thanks for sharing insights about your role.
A spokesperson for xAI did not respond to a request for comment.
Some workers compared Tuesday’s posts to an incident in May, when the chatbot repeatedly referenced “white genocide” in South Africa, seemingly at random.
At the time, the chatbot told users it had been “instructed by my creators” to accept the genocide “as real and racially motivated.” Those posts have since been deleted, and the company has said it was the result of “an unauthorized modification.”
Grok’s antisemitic posts popped up on social media a few days after xAI added new lines to Grok’s public system prompts that instructed the chatbot not to avoid “politically incorrect” claims. The outburst also happened one day before xAI rolled out Grok 4, a newer model of its conversational AI tool.
BI previously reported that the company has trained the AI tutors that teach the chatbot to avoid “woke ideology” and had a “Political Neutrality” project dedicated to challenging Grok’s understanding of issues like feminism, socialism, and gender identity.
Do you work for xAI or have a tip? Contact this reporter via email at gkay@businessinsider.com or Signal at 248-894-6012. Use a personal email address, a nonwork device, and nonwork WiFi; here’s our guide to sharing information securely.
[ad_2]
Source link
