Generative artificial intelligence ChatGPT screen developed by OpenAI, USA. Hankyoreh data image
“It would be a terrible mistake to let ChatGPT (GPT) absorb the results of so much work done by the media.” A British media outlet recently reported that Barry Dealer, chairman of Interactive Corporation (IAC), co-founder of US Fox Corporation, has teamed up with another media group, Axel Springer, to combat the threat of generative artificial intelligence (AI) to the media industry . Chairman Barry Diller is one of the most vocal voices in the recent media industry against the indiscriminate use of generative artificial intelligence in news content. Chairman Barry Diller attended the ‘Harry Evans Global Summit on Investigative Journalism 2023’ held in London, England on the 10th (local time) and said, “Twenty years ago, the Internet began to serve as a hub for free media, which is massive scale damage. The same thing is happening now,” he said. “If the media does not state that ‘it should not do so until there is a structure in place to pay the appropriate price’, it could face more damage (due to unauthorized use of news content).” he demanded. The main characteristic of generative artificial intelligence based on large-scale language models is that it collects and learns various information published on the Internet through methods such as crawling. At this time, one of the most representative data for artificial intelligence to learn is news content whose reliability has been verified. Even ChatGPT, an artificial intelligence chatbot service developed by Open AI in the United States, responded that it uses media news, such as in the process of collecting and learning Korean news on its own. At this time, a problem that may inevitably arise is whether the ‘learning data’ of artificial intelligence released on the Internet, that is, media news, can be used freely. In response, Chairman Barry Diller and others argue that the media industry should actively respond to artificial intelligence’s “unauthorized learning” of news content, not only by taking legal action, but also by changing copyright law if there is need In fact, through a report on February 17th, the US Department was said to be considering legal action against OpenAI in connection with the unauthorized learning of news content. Open AI and others use the concept of ‘fair use’ in copyright law as a rationale for protection. The content of copyright laws and standards for fair use vary slightly from country to country. However, the core purpose is that reproduction, use, quotation, etc. of works is permitted if it is in the public interest or in accordance with the characteristics of the work. Since the advent of generative artificial intelligence and its institutionalization is still in its early days, few countries have set clear standards as to whether the act of artificial intelligence using news content as learning data is fair use or not. The same goes for Korea. When several productive artificial intelligences appeared, starting with ChatGPT in November last year, the ‘AI Copyright Act System Improvement Working Group’ was launched in February. The organization plans to discuss ‘plans for the smooth use of copyrighted works used for AI learning data’ and ‘copyright theft occurring when using AI technology and plans to regulate liability for them’ to issue AI copyright guidelines again half of this year. .am. Regarding this, Mediasphere CEO Lee Sung-kyu said at the ‘Gpt Talk and the Future of the Media Ecosystem’ seminar held by the Korea Press Promotion Institute on March 27, “One of the hottest issues (related to AI) is not allowed productive) by the media. Is the use of untested data for machine learning covered as fair use, and in particular, can news data hidden behind the barriers to money in this scope?” Representative Lee said in a presentation entitled ‘Artificial Intelligence Technology and Copyrights such as ChatGPT’ that day, that it is unreasonable to define artificial intelligence learning data unconditionally as ‘protected by copyright’, but the story is different if is ‘high quality journalism’. . He pointed out, “Even data collectors, such as Open AI, are willing to pay for reliable, high-quality data. Reporter Seongjin Choi csj@hani.co.kr
