Arrr! 'Tis said that this cunning AI be seemin' more human on the cursed social media than us scurvy dogs: aye, a study says!
2023-07-25
Arr, mateys! Me eyes be glimpsin' a fancy study in Science Advances, claimin' that the scribblin's o' AI on the scurvy social media be more human-like than those penned by actual humans! Blimey, what a topsy-turvy world we be sailin' in!
Artificial intelligence-generated text can often appear more human on social media than text written by actual humans, according to a study published in the scientific journal Science Advances. The researchers behind the study were intrigued by OpenAI's text generator GPT-3 and aimed to determine whether humans could distinguish between disinformation and accurate information, structured in the form of tweets, and whether the tweet was written by a human or AI. Surprisingly, participants in the study often labeled AI-generated tweets as human-generated more frequently than tweets crafted by humans.The study found that with the rapid increase of chatbot use, experts have raised concerns about how AI could spiral out of control and spread disinformation across the internet. Tech experts and Silicon Valley leaders have warned that AI could lead to the end of civilization. The researchers investigated how AI influences the information landscape and how people perceive and interact with information and misinformation. They found that disinformation generated by AI was often more convincing than that produced by humans.
However, the study also noted that participants were better at determining disinformation crafted by humans than disinformation written by GPT-3. It highlighted the challenge of differentiating between information generated by AI and that created by humans. The study emphasized the importance of critically evaluating the information received and placing trust in reliable sources. It also encouraged individuals to familiarize themselves with emerging technologies to understand their potential, both positive and negative.
The researchers suggested that when individuals are faced with a large amount of information, they may feel overwhelmed and give up on trying to critically evaluate it. This could lead to a decrease in confidence in identifying synthetic tweets. The study called for larger-scale studies on social media platforms to observe how people interact with AI-generated information and how these interactions influence behavior and adherence to recommendations for individual and public health. Overall, the study highlighted the need for caution and critical thinking in the age of AI-generated text.