{"id":26022,"date":"2025-01-02T09:48:14","date_gmt":"2025-01-02T01:48:14","guid":{"rendered":"https:\/\/www.1ai.net\/?p=26022"},"modified":"2025-01-02T09:48:14","modified_gmt":"2025-01-02T01:48:14","slug":"%e4%bc%a6%e6%95%a6%e5%a4%a7%e5%ad%a6%e5%ad%a6%e9%99%a2%e3%80%81%e5%8d%8e%e7%9b%9b%e9%a1%bf%e5%a4%a7%e5%ad%a6%e7%a0%94%e7%a9%b6%ef%bc%9aai%e5%86%99%e4%bd%9c%e5%b7%a5%e5%85%b7%e6%ad%a3%e4%b8%8d%e6%96%ad","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/26022.html","title":{"rendered":"AI writing tools are improving, but still can't match human creativity: study by University College London, University of Washington"},"content":{"rendered":"

Today, people can let AI writing tools help them with just a few taps of the keyboardWriting essays, composing rap songs, or writing plays<\/strong>However, for the time being, mankind cannot expect it to have \"Shakespearean originality\". For now, however, humans can't expect \"Shakespearean originality\". According to Science, a new study suggests that these AI works are still clearly derivative, at least for now.<\/p>\n

\"97893b70j00spfuam002dd000sg00lcp\"<\/p>\n

To validate this idea, the researchers have developed a program that is capable ofMeasuring AI Creativity<\/strong>of the program.University of London<\/a>Mirco Musolesi, a computer scientist at the Academy who studies AI creativity, noted that assessing creativity is \"a complex and interesting challenge\" and argued that this new approach has been useful inAddressing linguistic novelty<\/strong>This aspect of the problem is outstanding.<\/p>\n

Since the introduction of generative AI and big language models, there have been questions in the scientific community about their creative capabilities. While these AIs can quickly generate text that appears to be human-created, some scholars argue that they are not truly innovative and simply reorganize the content in the training corpus. Critics have compared them to \"Random Parrot<\/strong>\" (Note: Blind repetition of known text).<\/p>\n

Yet quantifying this creativity is no easy task. Scientists usually take two approaches.Computerized detection of signs of plagiarism<\/strong>\u00a0-- but \"no plagiarism\" does not mean \"originality\"; and secondly, to let theHuman rating<\/strong>Assessing fluency and originality - but the processSubjective and time-consuming<\/strong>.<\/p>\n

In response.University of Washington<\/a>computer scientist Lu Ximing and his team have developed a tool called DJ Search. ItBoth objectivity and the ability to capture nuance<\/strong>. The tool extracts phrase fragments from the AI output text and looks for similar content in the database. It not only looks for exact matches, but also recognizes semantically similar expressions. With the help of AI algorithms to analyze word meanings, it determines synonyms through embedding vectors. Ultimately, it computesPercentage of unmatched portion in text<\/strong>to assess the novelty of the AI output.<\/p>\n

The study found that humans inIn the field of poetry, fiction and speech<\/strong>significantly outperforms AI, respectively.\u00a080%, 100% and 150%<\/strong>.<\/p>\n

DJ Search can also compare human productions, such as The Hunger Games, which has 35% more linguistic originality than Twilight. research leader Lu will\u00a0<\/strong>AI analogy for DJ<\/strong>: \"TheySplice text like a DJ mixing music<\/strong>The first is that, as wonderful as it is, it'sThere is no substitute for a composer.<\/strong>\u201d<\/p>\n

Nanyun Violet Peng, a computer scientist at the University of California, Los Angeles, suggests that the future should beAssessing the originality of the overall narrative<\/strong>, rather than being limited to the linguistic level.<\/p>","protected":false},"excerpt":{"rendered":"

Today, people can get AI writing tools to write essays, rap songs, and screenplays at the touch of a keyboard. For now, however, humans can't expect \"Shakespearean originality\". According to Science, a new study suggests that these AI works are still clearly derivative, at least for now. To test this idea, researchers have developed a program that measures AI creativity. Mirco Musolesi, a computer scientist at University College London who studies AI creativity, noted that evaluating creativity is \"a complex and interesting challenge\" and believes that the new approach does a good job of addressing the issue of linguistic novelty. Self-generating AI and big language modeling have been around for years.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[156,5364,1004],"collection":[],"class_list":["post-26022","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-5364","tag-1004"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/26022","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=26022"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/26022\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=26022"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=26022"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=26022"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=26022"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}