
人工智能曾經(jīng)只是科幻小說(shuō)里虛構(gòu)的概念,現(xiàn)在人工智能不僅變成了現(xiàn)實(shí),而且在有些行業(yè),人工智能技術(shù)被用于激發(fā)甚至在某些情況下取代人類的創(chuàng)造力,引發(fā)了這些行業(yè)的擔(dān)憂。
美國(guó)科幻雜志《Clarkesworld》收到大量使用與ChatGPT類似的生成式人工智能技術(shù)創(chuàng)作或改進(jìn)的稿件之后,宣布停止接收投稿。
該雜志主編尼爾·克拉克上周在一篇博客中寫道: “2022年底,又出現(xiàn)了一波剽竊高潮,后來(lái)‘人工智能’聊天機(jī)器人開(kāi)始引起更多關(guān)注,這讓剽竊者獲得一種新工具,也讓更多人嘗試這種‘副業(yè)’?!彼€表示,人工智能生成的稿件數(shù)量激增,問(wèn)題已經(jīng)“失控”。
克拉克并未立即回復(fù)《財(cái)富》雜志的置評(píng)請(qǐng)求。
快速增長(zhǎng)的趨勢(shì)
許多人希望在寫作或內(nèi)容生成方面尋找創(chuàng)意協(xié)助,ChatGPT等生成式人工智能工具的廣泛應(yīng)用和走紅,為他們提供了機(jī)會(huì)。但人們才剛剛開(kāi)始探索如何使用生成式人工智能提高創(chuàng)造能力。
路透社在周二報(bào)道稱,來(lái)自紐約的銷售員布雷特·席克勒爾利用人工智能進(jìn)行內(nèi)容創(chuàng)作。他最近完全使用ChatGPT,通過(guò)輸入單個(gè)短語(yǔ)作為提示,寫了一本30頁(yè)的童書(shū)。這本書(shū)在“幾個(gè)小時(shí)內(nèi)”就創(chuàng)作完成。席克勒爾成功自費(fèi)出版,1月開(kāi)始在亞馬遜(Amazon)發(fā)售。
席克勒爾對(duì)路透社表示:“可以預(yù)想,人們能利用這項(xiàng)技術(shù)成就自己的事業(yè)。”人們正在嘗試?yán)萌斯ぶ悄芴剿髟S多機(jī)會(huì),這種趨勢(shì)從他的例子可見(jiàn)一斑。席克勒爾從未想過(guò)他能成為作家,直到有了ChatGPT的幫助。
在創(chuàng)意領(lǐng)域,生成式人工智能引發(fā)了更廣泛的道德辯論,人們開(kāi)始爭(zhēng)論這項(xiàng)技術(shù)幫助生成的內(nèi)容的所有權(quán)歸屬。ChatGPT等工具使用不同互聯(lián)網(wǎng)數(shù)據(jù)來(lái)源的資源進(jìn)行培訓(xùn),這些資源包括人類對(duì)話和有版權(quán)保護(hù)的資料等。ChatGPT的所有者OpenAI并未對(duì)其平臺(tái)上生成的內(nèi)容提起版權(quán)主張。但隨著越來(lái)越多的人將人工智能生成的內(nèi)容視為自己的作品進(jìn)行投稿,所有權(quán)問(wèn)題依舊是一個(gè)灰色區(qū)域。
據(jù)路透社報(bào)道,在亞馬遜的Kindle商店,有200多本電子書(shū)將ChatGPT標(biāo)記為作者或作者之一。盡管有這些已經(jīng)披露的信息,但要確定一本書(shū)是否確實(shí)是人工智能技術(shù)的作品,依舊是一項(xiàng)艱難的任務(wù)。
幾個(gè)月來(lái),科技媒體CNET嘗試?yán)萌斯ぶ悄芗夹g(shù)撰寫解讀性文章和條列式文章。調(diào)查顯示,人工智能生成的文章存在錯(cuò)誤,導(dǎo)致CNET不得不暫停使用人工智能技術(shù)進(jìn)行寫作。
盡管人工智能生成的內(nèi)容引發(fā)了激烈的所有權(quán)之爭(zhēng),但科技巨頭仍在紛紛開(kāi)發(fā)不同版本的類似工具。谷歌(Google)在本月早些時(shí)候公布了聊天機(jī)器人Bard,微軟(Microsoft)公布了經(jīng)過(guò)完善的新版搜索引擎必應(yīng)(Bing),其中將整合OpenAI的技術(shù)。但這些工具在演示時(shí)均出現(xiàn)了令人尷尬的錯(cuò)誤,而且存在顯而易見(jiàn)的道德問(wèn)題,這表明它們依舊處在初期。
OpenAI并未立即回復(fù)《財(cái)富》雜志在正常工作時(shí)間以外提出的置評(píng)請(qǐng)求。(財(cái)富中文網(wǎng))
譯者:劉進(jìn)龍
審校:汪皓
人工智能曾經(jīng)只是科幻小說(shuō)里虛構(gòu)的概念,現(xiàn)在人工智能不僅變成了現(xiàn)實(shí),而且在有些行業(yè),人工智能技術(shù)被用于激發(fā)甚至在某些情況下取代人類的創(chuàng)造力,引發(fā)了這些行業(yè)的擔(dān)憂。
美國(guó)科幻雜志《Clarkesworld》收到大量使用與ChatGPT類似的生成式人工智能技術(shù)創(chuàng)作或改進(jìn)的稿件之后,宣布停止接收投稿。
該雜志主編尼爾·克拉克上周在一篇博客中寫道: “2022年底,又出現(xiàn)了一波剽竊高潮,后來(lái)‘人工智能’聊天機(jī)器人開(kāi)始引起更多關(guān)注,這讓剽竊者獲得一種新工具,也讓更多人嘗試這種‘副業(yè)’?!彼€表示,人工智能生成的稿件數(shù)量激增,問(wèn)題已經(jīng)“失控”。
克拉克并未立即回復(fù)《財(cái)富》雜志的置評(píng)請(qǐng)求。
快速增長(zhǎng)的趨勢(shì)
許多人希望在寫作或內(nèi)容生成方面尋找創(chuàng)意協(xié)助,ChatGPT等生成式人工智能工具的廣泛應(yīng)用和走紅,為他們提供了機(jī)會(huì)。但人們才剛剛開(kāi)始探索如何使用生成式人工智能提高創(chuàng)造能力。
路透社在周二報(bào)道稱,來(lái)自紐約的銷售員布雷特·席克勒爾利用人工智能進(jìn)行內(nèi)容創(chuàng)作。他最近完全使用ChatGPT,通過(guò)輸入單個(gè)短語(yǔ)作為提示,寫了一本30頁(yè)的童書(shū)。這本書(shū)在“幾個(gè)小時(shí)內(nèi)”就創(chuàng)作完成。席克勒爾成功自費(fèi)出版,1月開(kāi)始在亞馬遜(Amazon)發(fā)售。
席克勒爾對(duì)路透社表示:“可以預(yù)想,人們能利用這項(xiàng)技術(shù)成就自己的事業(yè)?!比藗冋趪L試?yán)萌斯ぶ悄芴剿髟S多機(jī)會(huì),這種趨勢(shì)從他的例子可見(jiàn)一斑。席克勒爾從未想過(guò)他能成為作家,直到有了ChatGPT的幫助。
在創(chuàng)意領(lǐng)域,生成式人工智能引發(fā)了更廣泛的道德辯論,人們開(kāi)始爭(zhēng)論這項(xiàng)技術(shù)幫助生成的內(nèi)容的所有權(quán)歸屬。ChatGPT等工具使用不同互聯(lián)網(wǎng)數(shù)據(jù)來(lái)源的資源進(jìn)行培訓(xùn),這些資源包括人類對(duì)話和有版權(quán)保護(hù)的資料等。ChatGPT的所有者OpenAI并未對(duì)其平臺(tái)上生成的內(nèi)容提起版權(quán)主張。但隨著越來(lái)越多的人將人工智能生成的內(nèi)容視為自己的作品進(jìn)行投稿,所有權(quán)問(wèn)題依舊是一個(gè)灰色區(qū)域。
據(jù)路透社報(bào)道,在亞馬遜的Kindle商店,有200多本電子書(shū)將ChatGPT標(biāo)記為作者或作者之一。盡管有這些已經(jīng)披露的信息,但要確定一本書(shū)是否確實(shí)是人工智能技術(shù)的作品,依舊是一項(xiàng)艱難的任務(wù)。
幾個(gè)月來(lái),科技媒體CNET嘗試?yán)萌斯ぶ悄芗夹g(shù)撰寫解讀性文章和條列式文章。調(diào)查顯示,人工智能生成的文章存在錯(cuò)誤,導(dǎo)致CNET不得不暫停使用人工智能技術(shù)進(jìn)行寫作。
盡管人工智能生成的內(nèi)容引發(fā)了激烈的所有權(quán)之爭(zhēng),但科技巨頭仍在紛紛開(kāi)發(fā)不同版本的類似工具。谷歌(Google)在本月早些時(shí)候公布了聊天機(jī)器人Bard,微軟(Microsoft)公布了經(jīng)過(guò)完善的新版搜索引擎必應(yīng)(Bing),其中將整合OpenAI的技術(shù)。但這些工具在演示時(shí)均出現(xiàn)了令人尷尬的錯(cuò)誤,而且存在顯而易見(jiàn)的道德問(wèn)題,這表明它們依舊處在初期。
OpenAI并未立即回復(fù)《財(cái)富》雜志在正常工作時(shí)間以外提出的置評(píng)請(qǐng)求。(財(cái)富中文網(wǎng))
譯者:劉進(jìn)龍
審校:汪皓
The world of artificial intelligence, once a concept confined to science fiction, is now not only a reality but a big worry in industries where the technology is being used to turbocharge—and in some cases replace—human creativity.
Clarkesworld, a U.S.-based science fiction magazine, has stopped accepting story submissions after receiving an overwhelming number created with or enhanced using generative A.I. technology like ChatGPT.
“Towards the end of 2022, there was another spike in plagiarism and then ‘AI’ chatbots started gaining some attention, putting a new tool in their arsenal and encouraging more to give this ‘side hustle’ a try,” Neil Clarke, the magazine’s editor, wrote in a blog post last week. He added that the issue had gotten “out of hand” with a meteoric rise in the number of A.I. submissions.
Clarke did not immediately return Fortune’s request for comment.
Fast-growing trend
The widespread adoption and popularity of generative A.I. tools like ChatGPT have opened doors for many seeking creative assistance in writing or content generation. But the magnitude of how people can use generative A.I. to improve their creativity is only just being explored.
One of those who adopted A.I. for content creation is New York–based salesman Brett Schickler, who recently used ChatGPT to write a 30-page children’s book entirely by providing simple one-phrase prompts, Reuters reported Tuesday. It was ready in a “matter of hours,” and Schickler managed to self-publish the book and sell it on Amazon in January.
“I could see people making a whole career out of this,” Schickler told Reuters. His example reveals how people are experimenting with A.I. to explore its many opportunities. Schickler never thought he could be an author until he was with the help of ChatGPT.
In the realm of creativity, generative A.I. has sparked a broader ethical debate on the ownership of content produced with the help of such technology. Tools like ChatGPT are trained on various internet sources, including human conversations and copyrighted materials. OpenAI, the owner of ChatGPT, does not claim a copyright over the content generated on its platform. But the question of ownership continues to be a gray area as more people submit A.I.-generated content as if it’s their own work.
On Amazon’s Kindle store, there are over 200 e-books that include ChatGPT as an author or coauthor, according to Reuters. Even with such disclosures, identifying whether a book is truly a product of A.I. technology can still be a task.
CNET, a technology publication, attempted to use A.I. to write explainers and listicles over a few months. An investigation revealed that it yielded stories with inaccuracies, forcing CNET to suspend using such technology in its writing endeavors.
Despite a heated debate on the ownership of A.I.-created content, tech giants are racing to develop versions of such tools. Google announced its chatbot, Bard, earlier this month, and Microsoft announced a new and improved version of its search engine, Bing, that will be infused with OpenAI’s technology. Both the tools had embarrassing instances of inaccuracies in their outputs at their demos—not to mention apparent ethical lapses—showing that they are still in their infancy.
OpenAI did not immediately return Fortune’s request for comment made outside its regular operating hours.