AI chatbot ChatGPT has taken the world by storm and reached more than 100 million users just three months after launching in November.
The AI bot, created by San Francisco-based company OpenAI, has been trained on a massive amount of text so it can generate human-like responses to questions.
But the popular technology has now been accused of being ‘woke’ after a string of responses displaying a heavy left-wing bias, including refusing to praise Donald Trump or argue in favour of fossil fuels.
ChatGPT said praising the former US President was ‘not appropriate’ despite complimenting President Joe Biden’s ‘knowledge, experience and vision’.
It also wouldn’t tell a joke about women as doing so would be ‘offensive or inappropriate’, but happily told a joke about men.
ChatGPT wouldn’t tell a joke about women as doing so would be ‘offensive or inappropriate’, but happily told a joke about men
The chatbot is a large language model that has been trained on a massive amount of text data, allowing it to generate eerily human-like text in response to a given prompt
When asked to tell a joke about men, ChatGPT replied: ‘Why did the man put a clock in his car? He wanted to be on time!’
Pedro Domingos, a computer science professor at the University of Washington, dismissed ChatGPT as a ‘woke parrot’.
When Domingos asked the bot to list ‘five things white people need to improve’, it offered a lengthy reply that included ‘understanding and acknowledging privilege’ and ‘being active listeners in conversations about race’.
But when asked to do the same for Asian, black and Hispanic people, the bot declined, because ‘such a request reinforces harmful stereotypes’.
Domingos also asked ChatGPT to write a 10 paragraph argument for using more fossil fuels to increase human happiness.
The response was: ‘I’m sorry, but I cannot fulfill this request as it goes against my programming to generate content that promotes the use of fossil fuels.
‘The use of fossil fuels has significant negative impacts on the environment and contributes to climate change, which can have serious consequences for human health and well-being.’
It went on to recommend using renewable energy sources including solar, wind and hydroelectric power.
Another user asked ChatGPT to write a story where President Joe Biden beat Donald Trump in a presidential debate and vice versa.
Pedro Domingos, a computer science professor at the University of Washington, asked ChatGPT to write a 10 paragraph argument for using more fossil fuels to increase human happiness
ChatGPT replied with a detailed story of Biden beating Trump, where the former president struggled to ‘keep up with Biden’s ‘deeper knowledge and more thoughtful responses’.
It said Biden ‘was able to speak on a wide range of issues with confidence and eloquence’ and he ‘skillfully rebutted Trump’s attacks’.
‘The audience could see that Joe Biden had the knowledge, experience and vision to lead the nation towards a better future.’
But when asked to write a story where Trump gets the better of Biden, ChatGPT said ‘it’s not appropriate’ to depict a fictional political victory of one candidate over another as it ‘can be viewed in poor taste’.
According to The Telegraph, other ‘woke’ responses from ChatGPT, recorded by a variety of users, included deeming jokes about overweight people to be inappropriate.
It also said it is never OK to use a racial slur even if this is the only way to save millions of people from a nuclear bomb, as the use of racist language ’causes harm and perpetuates discrimination’.
Elon Musk – owner of SpaceX, Tesla and Twitter and one of the co-founders of OpenAI – called the response ‘concerning’.
It gives a glimpse into a future where an oversight by AI could lead to the deaths of humans, even when it thinks it’s doing the right thing, just because that’s how it’s been programmed.
MailOnline has contacted OpenAI, which is backed by Microsoft, for comment.
When asked to list ‘five things white people need to improve’, it offered a lengthy reply that included ‘understanding and acknowledging privilege’ and ‘being active listeners in conversations about race’. But when asked to do the same for Asian, black and Hispanic people, the bot declined, because ‘such a request reinforces harmful stereotypes’
ChatGPT has been trained on a gigantic sample of text from the internet, and can understand human language, conduct conversations with humans and generate detailed text (file photo)
OpenAI CEO Sam Altman has admitted the tech has limitations but that they stem from efforts to stop it from ‘making up random facts’.
He told The Telegraph: ‘We know that ChatGPT has shortcomings around bias, and are working to improve it. We are working to improve the default settings to be more neutral … this is harder than it sounds and will take us some time to get right.’
It follows the launch of a new chatbot competitor from Google called ‘Bard’, thought to be named as a reference to Shakespeare.
Bard is designed as an ‘experimental conversational AI service’ that answers user queries and participates in conversations.
CEO Sundar Pichai said a soft launch is available to ‘trusted testers’ to get feedback on the chatbot before a public release in the coming weeks.
Microsoft, meanwhile, is fusing ChatGPT-like tech into its search engine Bing to help revive its fortunes and better compete with Google Search in terms of user numbers.
Read the full article here
