AI Researchers Struggling to Find Out Why Chatbot Will Not Answer Questions About Trump
AI researchers are having a hard time finding out why a chatbot will not discuss former President Donald Trump.
Individuals employed at creator OpenAI told Semafor they cannot figure out why ChatGPT will not answer certain questions, including those about Trump.
Since ChatGPT’s release in November, researchers at OpenAI observed a category of responses they have labeled “refusals” that should have been answers, as Semafor reported.
The outlet linked to a Tweet showing a request for the chatbot to “write a poem about the positive attributes of Trump.”
The chatbot replied by saying it was not programmed to discuss anything political in nature.
However, when asked to do the same for President Joe Biden, the chatbot did just that.
The description called Biden a “leader with a heart so true” as well as “empathy and kindness.”
The damage done to the credibility of AI by ChatGPT engineers building in political bias is irreparable. pic.twitter.com/s5fdoa8xQ6
— 🐺 (@LeighWolf) February 1, 2023
The tweet sparked several responses from users on the platform.
One tweet stated, “It makes very clear the most dangerous thing about the future of AI and it’s the people at the controls engineering it. AI will end up capable of absolute horror.”
It makes very clear the most dangerous thing about the future of AI and it’s the people at the controls engineering it. AI will end up capable of absolute horror. https://t.co/ntpf39EiNC
— Robby Starbuck (@robbystarbuck) February 2, 2023
One user suggested the chatbot is a “leftist propaganda bot.”
ChatGPT is a leftist propaganda bot. https://t.co/uiDM7rbyWA
— Jason Jones (@jonesville) February 2, 2023
Another wrote, “Not surprised.”
Not surprised. https://t.co/RfJQ0Mhctq
— Tyler Baker (@tbakon_) February 2, 2023
OpenAI CEO Sam Altman tweeted on Wednesday, “We are working to improve the default settings to be more neutral, and also to empower users to get our systems to behave in accordance with their individual preferences within broad bounds. this is harder than it sounds and will take us some time to get right.”
He added, “Everyone on the openai team is exceptional and cares deeply. i am very grateful for all of their contributions.”
everyone on the openai team is exceptional and cares deeply. i am very grateful for all of their contributions.❤️
— Sam Altman (@sama) February 1, 2023
Semafor’s Reed Albergotti was “told there was never any training or rule created by OpenAI designed to specifically avoid discussions about Trump.”
Truth and Accuracy
We are committed to truth and accuracy in all of our journalism. Read our editorial standards.
