NewsGPT, the world’s first AI-generated news site, has been challenging the very codes of journalistic conduct. For too long, news channels have been plagued by bias and subjective reporting, Alan Levy, CEO at NewsGPT, said of his motivation to launch a news channel completely driven by AI.
Unlike other news channels, the news on this portal is not influenced by advertisers, political affiliations, or personal opinions, a press release said. While it’s a noble thought from Levy, the approach has the potential to be catastrophic.
LLMs are Hallucinatory
NewsGPT has not revealed much besides they use a combination of AI programmes. However, as the name suggests, Levy’s firm could be leveraging a version of OpenAI’s GPT Large Language Models (LLM). ChatGPT, the popular chatbot by OpenAI, is based on the GPT3.5 architecture.
AIM Daily XO
Join our editors every weekday evening as they steer you through the most significant news of the day, introduce you to fresh perspectives, and provide unexpected moments of joy
Your newsletter subscriptions are subject to AIM Privacy Policy and Terms and Conditions.
Time and time again, various AI experts have pointed out that current LLMs are hallucinatory in nature. When ChatGPT caught the world’s attention, Sam Atlman, one of the co-founder at OpenAI, pointed out that it would be a mistake to be relying on it for anything important right now. “It’s a preview of progress and we have lots of work to do on robustness and truthfulness.”
Yann LeCun, chief AI scientist at Meta, raised the same issue. According to him, such models have no knowledge of the world around them. “They have no idea of physical reality. They don’t have any context and they can’t find their answers,” he said, at the World Artificial Intelligence Cannes Festival.
Download our Mobile App
The accuracy levels of these models remains significantly poor, hence completely relying on them, for something as significant as journalism, could lead to catastrophic consequences.
Interestingly, this won’t be the first time that a media house is leveraging AI. Earlier this year, US-based tech publication CNET announced that they are using an internally designed AI model to churn out articles. However, the content generated by CNET’s AI were found to be riddled with factual errors.
Breach of Journalistic Values
The Press Council of India (PCI) states ‘accuracy and fairness’ as one of the first principles and ethics for journalistic conduct. With what we have seen so far with CNET, AI-generated content fails miserably in terms of accuracy and fairness.
Normally, in a news organisation, a story goes through various layers of checks, are vetted and fact-checked by editors before being published. In fact, checking the factual accuracy of the story is another core journalistic value.
In the case of NewsGPT, there are no humans involved in the whole process. The company has confirmed to Futurism that “There are no human fact-checkers. Our news stories are generated 100 percent by AI.”
Further, there is also the problem of AI’s political inclination. ChatGPT, the popular chatbot by OpenAI, was found to be politically biassed. Researchers found that the chatbot showed pro-environmental, left-libertarian ideology. “The information and data fed into AI models can reflect societal and cultural biases, leading to biassed results in the predictions made by the AI model,” Salman Waris, partner head of TMT & IP Practice at TechLegis Advocates and Solicitors, told AIM.
While Levy claims his news reporting AI is not influenced by political affiliations, the underlying model on which NewsGPT has been built on, could be inherently biassed. Moreover, while NewsGPT is the portal in question today, tomorrow it could be another platform that publishes biassed content. Hence, PCI, as well as other adjudicating organisations worldwide, must take cognizance of this development.
AI as an Enabler
AI, in its current form, is not sophisticated enough to take over the media industry. In fact, it could be a great tool for journalists. We at Analytics India Magazine, have been using numerous AI tools such as MidJourney and ChatGPT for quite some time. “I personally use the AI-powered tool DALL.E to produce images for my stories,” Imtiaz Shariff, news editor at Deccan Herald, said.
Digital media platform The Better India, has previously told AIM that it leverages ChatGPT for its social media marketing efforts.
