The appeal presented to the British Parliament on the difficulties related to the training of artificial intelligence and Japan’s decision to make an exception on the validity of copyright in applications of artificial intelligence demonstrate that the future of copyright and Copyright is becoming increasingly nebulous, says Kristjan Port in The R2 Technical Commentary.
Humanity is faced with an important choice. If you want something new, you have to work hard for it, think furiously and act energetically, doing everything differently than usual and outside the norms that the majority adheres to. A creative or scientific idea would fit any brain, but this quality consists not in the intellectual drive of the masses, but in the individual capture of fantasies. It may happen once in a lifetime, or at least it may take a long time before you encounter it.
To motivate a part of society to stay on the indicated difficult path, a reward scheme was created, which guarantees those who find a witty idea the opportunity to profit from it. They started talking about intellectual property and copyright. Over time, the system has been complemented by more indirect protection of owners’ interests, including the right to make copies
What matters for society is still the final result, not individual economic performance. Only society is the channel of economic benefits. No person can guarantee this. Consent and social protection are necessary. By giving rights to an individual, society protects its interests.
The new dilemma is whether to continue with the old model of the human-centered creator or place our hopes in creative AI. Believing in the latter, one must cancel part of the previous agreements. Otherwise, current copyright and authorship holders will prohibit or at least prevent artificial intelligence from learning about human creations and learning the creative process.
No one knows what AI will ultimately be capable of. Considering current achievements and the speed of development, there are certainly many who can tentatively be called techno-optimists. It is not at all difficult to gather an army of people with the opposite opinion, who talk about how important it is to have heart and soul in the creative process, to feel pain, etc. This makes the decision difficult, because everyone wants a better and more creatively rich future.
The beginning of the year brings this consideration forcefully onto the agenda. OpenAI, the artificial intelligence developer, just admitted that artificial intelligence with extraordinary creative capabilities, such as the GPT-4 model, cannot be created without learning from human creations. This means that it is impossible to develop a new generation of artificial intelligence without using copyrighted materials.
A paper presented to the House of Lords of the British Parliament points out that virtually all modern human expression, from online posts to government documents, is in some way protected by property rights. It would be essentially impossible to develop next-generation AI models without using such protected materials. Consequently, hoping for the development of a new generation of technology, it is necessary to make concessions to these rights.
Considering that around 95% of people read and watch and that the remaining 5% generate a significant portion of the content, the interest and power are in the hands of the majority. At the same time, they themselves are mostly not aware of this. It is by no means a comfortable power, the incorrect application of which pays dearly and lacks the hoped-for creative fruits.
The situation is exemplified by recent and likely further court disputes in the near future, as well as Japan’s official position. First of all, the legal disputes, where, for example, one of the main US newspapers, the New York Times, accuses OpenAI of learning without authorization from materials created by journalists. More subtly, this reaction reflects fears about a text generation service competing with journalism. At the same time, it is clear that the owners of the newspaper consider the OpenAI language model to be an important competitor.
Until now, everyone has also been allowed to memorize newspaper articles without endangering the press. Some news organizations already use language models instead of journalists. This in turn provokes a storm of indignation, as happened with the well-known sports magazine Sports Illustrated, where readers were offered both images and reading materials of artificial origin.
Copyright infringement charges have also been filed against several AI companies that produce visual creations. By many estimates, a bigger conflict lies ahead in the world of music. Actors and screenwriters in the United States have just ended a six-month strike, prompted by fears that movie studios will replace them with creations of artificial intelligence, etc. These cases also highlight that artificial intelligence has a creative potential at least comparable to that of humans. The same trust is demonstrated by the investments made in the development of the sector.
At the same time, the criticisms concern technologies that have only been in use for a few years. While it is in everyone’s interest to bring about a new era of prosperity and creative quality, there is growing pressure to review existing goals and agreements on intellectual property protection.
As an example of the new era, Japan, the world’s third-largest economy, has decided that copyright restrictions will not extend to teaching artificial intelligence. The Japanese government is of the opinion that the laws that concern it hinder the development of an important sector for the country.
From Monday to Thursday you can listen to Kristjan Port’s technological commentary on Radio 2’s “Portaal”.
2024-01-09 08:52:00
the-formation-of-artificial-fur-calls-into-question-the-future-of-copyright-r2-portal