Artificial intelligence (AI) itself does not spread misinformation, says automated text generator ChatGPT. According to it, AI algorithms can be developed to automate the generation and distribution of false information, but it is humans who design and deploy these algorithms for malicious purposes.
This advanced AI language model, designed to replicate human conversations, was launched in November 2022 and according to the latest available data, currently has more than 100 million users, making it the fastest-growing app in history. Just for comparison, it took TikTok 9 months to reach 100 million users and Instagram 2.5 years.
The chatbot developed by OpenAI has sparked a lot of controversy and challenges. Whether it's a homework assignment, an essay, a journalistic article, or a more in-depth research, ChatGPT is "loaded" to answer any number of questions that anyone can ask it. Where the answers end up is hard to find and even harder to check for accuracy.
Artificial intelligence puts journalism in an unenviable position
Experts in this field estimate that automatic text generators are capable of creating any text that cannot be recognized whether written by a human being or a machine. According to them, ChatGPT, by “seeing” and “learning” from what journalists have written so far, faces the real possibility of overproducing texts with, on average, the same quality as those written so far by journalists.
Scholars of this issue see ChatGPT's role in journalism as both positive and negative. On the positive side, this technology can help journalists fill in those parts of journalistic content where the story of the event being written about needs to be briefly explained. According to them, human judgment, critical thinking, and contextual understanding remain critical in interpreting and presenting news accurately and ethically. Artificial intelligence should be seen as a tool to enhance journalistic skills, supporting journalists in their work, rather than replacing them.
The line between disinformation and automatically generated text is thin
According to experts, the line between disinformation and automatically generated text is thin, and that disinformation cannot be attributed to AI. ChatGPT acknowledges that AI can be susceptible to disinformation and recommends several factors to consider. Such as the fact that AI models like ChatGPT learn from large amounts of data, including texts from the internet, so if that data contains biased or inaccurate information, the model could inadvertently spread that misinformation in its responses.
To mitigate these challenges, ChatGPT recommends incorporating human oversight and fact-checking processes when using AI-generated content.
According to ChatGPT, preventing the sharing and creation of fake news with artificial intelligence (AI) requires a comprehensive and multifaceted approach that includes legal, technological, and ethical considerations.
You can find the full article at mim.org.mk