![Launch your website in 2025 with Hostinger at up to 80% lower cost [+ an extra 10%]](https://res.cloudinary.com/dcj0jkqds/image/upload/v1736275898/posts_previews/cukixmumxgk7zb5qur6o.png)
Can you imagine Neruda's opinion? Poetry generated by artificial intelligence surpasses that written by humans.
A study conducted by the University of Pittsburgh revealed that the ability to discern the authorship of poetry is affected in humans, who are unable to differentiate between works written by artificial intelligence and those created by people.
A recent study conducted by a team from the University of Pittsburgh has demonstrated that artificial intelligence has the capacity to create poetry that is indistinguishable from that written by humans. In fact, participants in the research showed a preference for AI-generated poems compared to works by iconic poets like Shakespeare and Dickinson. In an experiment with over 1,600 readers, they were only able to correctly identify whether a poem was produced by a machine or a person 46.6% of the time.
The results revealed that poems created by AI received higher scores in qualitative characteristics such as rhythm, beauty, and emotional impact compared to texts by recognized poets. Interestingly, five poems considered the least human belonged to famous poets, while four of the poems deemed most human were generated by artificial intelligence. This finding could cause discomfort in some sectors of the literary community, as it indicates that the distinction between human and AI-generated poetry is becoming increasingly complicated. Furthermore, the study highlights that AI could surpass humans in surprising aspects of creative culture.
In another area, Google reported that its tool for detecting AI-generated content, known as SynthID, has been released for public use. This system introduces invisible watermarks in images, videos, and texts generated by AI, allowing users to authenticate the provenance of the content. The company emphasized that this tool will be accessible for free to developers and businesses, facilitating the identification of machine-generated content.
Meanwhile, Canva has launched a new image generation tool called "Dream Lab," which allows users to create images from descriptions in different styles, thanks to the incorporation of the generative AI startup Leonardo.ai. In addition, Canva has enhanced its set of AI tools known as "Magic," which now includes new functionalities such as automatic subtitle generation and animation effects for videos.
Regarding robotics, researchers from Penn Engineering have pointed out that current language models used in robotic platforms are not secure enough when interacting with physical environments. This discovery underscores the need to strengthen the safety of AI-based systems operating in the real world, given the potential dangers they may pose.
