The INNOV8 Project – Intellectual Property Education for Young Innovators

Every time artists play a piece of music, they are granted copyright. This is a subject of law that protects the author of a work and determines how he may profit from it and control its dissemination.

Copyright protects any musical work of a creative nature and does so automatically, at the very moment of creation.

In particular, copyright goes by default to the author of the creative work, regardless of who he works for or whether he wants to exercise this right.

There is a whole set of rights that belong to the artist. But what are they?

They are called patrimonial rights, i.e. those that provide an economic benefit to the author or the person who owns them if the author has assigned the rights to someone else. In this group, there are also moral rights, i.e. a set of rights that include the so-called right of authorship of the work, which concerns the link between the person and his work.

Even after his death, an author is recognized as the father (or mother) of his/ her work. When an artist uploads their song on streaming platforms, royalties come into play.

In general, the payments are made to the owners of a particular musical work, recording, melody, or performance.

There are two types of music royalties: performance royalties and mechanical royalties.

The first one relates to all performances, occasions when the song is played on the radio, on TV, in a presentation, in a video game, in a shop, or in a live venue. The latter is related to the sale of physical copies of the music, downloads, and streaming.

In the second half of 2022, 612 million people had subscribed to streaming music platforms. In 2022, streaming revenues reached $17.5 billion worldwide, the highest figure ever recorded.

Since 2015, the value of this market has increased sixfold. In 2023, in Italy alone, digital music market revenues reached almost USD 198 million.

The income of artists is linked to royalties, and for Italian singers on audio streaming platforms, it happens through contractual agreements between the artists and the record labels or music publishers. When I put a song on a platform like Spotify or Apple Podcast, the platform itself keeps a percentage for itself and another part is used to pay the artists. Except that the payment is usually not direct, but is made through record labels or distributors who in turn have a contract with the artist.

First, we have to go back to royalties: artists receive a percentage of the revenue generated by the streaming platform.

This percentage is usually negotiated between the artist and the record label or publisher when the contract is signed and therefore varies depending on the type of contract, and the notoriety of the artist. Royalties are then divided among the various parties that have rights to the music, i.e. the singer, composer, producer, and any others involved in its creation. This division is also decided when the contract is signed and according to each party’s share.

Then there is the question of the timing of payments: streaming platforms generally make periodic payments to artists. Payments are usually based on the number of plays or the artist’s market share.

Recently, a bit of a case has exploded over what payments artists receive.

The algorithms for calculating artist payments are said to be opaque and unclear. According to some, the distribution of revenues is not entirely fair, so there have been efforts to improve clarity in the financial relationship between platforms and artists. For a stream to be valid for payment on Spotify, the song must be played for at least 30 seconds.

Between 2016 and 2021, artists’ remuneration increased by 96% and record companies’ by 63%.

Only 1,060 of the 8 million artists on Spotify have earned more than one million dollars from streaming their tracks on this platform. This is a very low number, 0.01% of the total. We have the theory, but in practice, how much do artists earn from the rights they have to their songs if we listen to them on streaming platforms?

The answer to this question is not easy. There was one case this year, however, that was emblematic. First of all, it must be said that it is a US case, in a country that has different laws from ours.

It is the story of ‘Weird Al’ Yankovic, an American singer-songwriter who in November 2023 posted a video to thank his listeners on Spotify. What was the problem, however? Yankovic claimed that with his 80 million streams on the platform, he would earn a beauty of…$12. According to Producer Hive, Spotify would pay artists a figure of $0.003 or $0.004 per play. In short, 80 million streams should correspond to about 320,000 euros, a far cry from what Yankovic claimed.

Who is right? Hard to say. A lot depends on the agreements made between record companies, singers, and authors and the division of royalties.

What is certain is that Spotify has stated that from the beginning of 2024, songs will have to have reached at least 1,000 plays in the previous 12 months to generate royalties. The company claims that this change will help very active musicians increase their earnings. The idea is that millions of tracks have listened to less than 1,000 and therefore realize very small payments, but at the same time there is a bank commission on each of these, which often results in the payments due to the artists ending up in nothing. At the same time, many argue that this decision by Spotify may further penalize emerging musicians while valuing those who are more established in the industry.

Italian consumers spend on average 20 hours per week listening to music from streaming platforms

Our country has shown that it wants to safeguard the economic interests of artists. Italian law states that authors and performers, when they grant licenses to use their works or assign exclusive rights for exploitation, are entitled to receive a payment in line with the value of the rights granted.

The payment must be commensurate with the revenue that the licensor makes from that work. For example: I agree with a streaming platform to publish one of my songs. Here, the payments I am due must be commensurate with how much money the platform itself makes through my song.

In a study published in November 2023 by the European Parliament, it was found that most artists and their representatives at the European level believe that there is a strong and urgent need for a law protecting remuneration per song that is commensurate with the licensing and exploitation rights transferred. In particular, the longer it takes to have a law, the more difficult it will be to eradicate the bad practices already in place. On the contrary, producers and broadcasters do not take the same approach and even believe that no changes are necessary.

3.2 billion photos and 720,000 hours of videos are uploaded to social media every day. But when we upload pictures or images on the internet and social media, we often forget one crucial aspect: who owns the rights to those images?

It is important to understand this to avoid legal problems, respect the work of others, and protect our photos.

When we create an image or photo, according to the law, we automatically own the copyright, and no one can take and use it without our consent. What if we share it online? This is where licenses come into play.

A license states how other people can use your image.

There are more restrictive licenses that restrict use to personal purposes only, and more open licenses that also allow commercial uses.

Let’s take an example: Creative Commons licensed images. Creative Commons licenses are contracts whereby the copyright holder permits others to use the work. But this can be done with limitations. One can freely choose which rights to grant for use and which to keep reserved. This is why Creative Commons licenses are divided into certain types:

• some allow us to use the photos, with the sole constraint of indicating who the author is;
• then some require us not to make commercial use of the image, i.e. they cannot be used for activities that make us money;
• those require us to use the image exactly as it is, without modifying it, or there may be mixes of all the above.

The only images that we can use without any problem are those that have the ‘public domain’ indication.

But if you have any doubts about these licenses, the Creative Commons site gives all the specifications.

85% of photos uploaded online are ‘stolen’, i.e. used without the rights holder’s consent.

Italy ranks third in Europe for image rights violations, after Germany and the UK. Posting photos or videos is one of the activities Italians do most on social media, preceded only by liking other users’ posts and sending private messages.

There is a step, precisely about the licensing of images, that is crucial to take into account. The moment we upload photos online, on our Instagram profile, for example, we accept the platform’s Terms of Service.

These documents contain certain clauses regarding the transfer of rights, just like licenses. In practice, when we upload an image to a social media platform, we assign some rights to the company operating the platform. Instagram in particular states that it is committed to helping people and organizations protect their intellectual property rights, and that its terms of use do not allow content to be posted that infringes another person’s intellectual property rights, including copyright. This means that if someone takes a photo of us posted there without our consent, it not only violates Italian law but also the platform’s internal rules.

TikTok also works similarly. It does not allow content that infringes or violates copyright to be published, shared, or posted on the platform, subject to exceptions.

EU law provides a list of exceptions to copyright granted to right holders, each with a specific scope.

For instance, a teacher can use educational content such as an image or video posted online to deliver his or her lesson without infringing anyone’s copyright.

According to some estimates, if we were to read Instagram’s Terms of Service in full, it would take us 9 minutes and 42 seconds, TikTok 31 minutes and 24 seconds, and Microsoft even 1 hour and 3 minutes.

Exceptions to copyright infringement are imposed country by country by law. All platforms must respect them.

97% of 18-34 year olds accept the Terms and Conditions while using online services without reading them. As of 2022, the Digital Service Act in the EU recognizes social platforms as ultimately responsible for publishing harmful or illegal content. They are obliged to improve content moderation, according to the principle that what is illegal offline is also illegal online.

So is stealing an image like stealing a wallet?

Let’s start with the wallet. In Italy, if I steal a wallet, it is called theft. Theft is governed by Article 624 of the Criminal Code, with penalties varying according to the value of the stolen wallet.

But how does it work for digital content? There are cases, not at all rare, in which the ‘theft’ of photographs published online can become a criminal offense.

The first case is provided for in Article 171 L.A., which punishes anyone who reproduces or sells a work resulting from the creativity of another person, without having the right to do so. In these cases, only the creative work, i.e. the creative photograph, is criminally protected.

Then there is the case of generic photographic images, i.e. those photos that reproduce realities, persons, objects, and elements of private or social life, but without the characteristics of a creative work. Here the venue is civil and not criminal.

In this case, compensation can be claimed in favor of the owner of the rights to the image, by the person who has abused it. In addition, the author of the image can request its immediate removal from the online channels where it was published.

However, when does this theft become criminally relevant? Meanwhile, anyone who publishes images of other people without their consent and intending to profit for himself or others, or to cause harm to the person concerned, commits the offence of unlawful processing of personal data.

If our image is then used to create a fake profile, the person who uses our photo commits an offense not so much related to copyright, but that of ‘substitution of person’ (Art. 494 of the Criminal Code) and is punishable by imprisonment of up to 1 year. Our identity can be damaged and our reputation affected.

In 2023, eight people were arrested in Italy for online fraud that committed the crime of impersonation.

According to the ‘Censis-DeepCyber’ report, Facebook suffered three different cyber attacks between 2018 and 2022, which resulted in the violation of the personal data of more than 500 million members (35 million Italians were reportedly affected).

We understand what we risk if we steal a photo, but what can we do if someone has stolen and used one of our pictures?

First, we need to find out exactly where our photo was published. To do this, there are reverse search engines or specific apps to find all copies of our image online. It can also help to contact a lawyer or legal advisor who specializes in digital rights.

If we discover that our image has been published on a specific social network or site, we should contact them. Many platforms have procedures for removing unauthorized content that allow us to report the author of the theft to the platforms. It is always advisable then to refer to the site of the postal police to report the violation, which, as we have seen, could be a crime.

And finally, to protect ourselves, we can consider updating our privacy settings on social media. By restricting access to our content, we can have more control over who can share and download our images.

The generative artificial intelligence market is expected to grow from USD 67 billion in 2023 to over USD 1.3 trillion in 2032.

First of all: what is generative artificial intelligence? It is a very powerful frontier in artificial intelligence research and development. Unlike the most basic AI tools, this type of artificial intelligence can generate (precisely) original and creative content, such as images, text, or even music. The creation of images in particular is one of the most surprising and fascinating results of this technology.

But how does it work? The idea, simplifying, is this: as with all other artificial bits of intelligence, the basic structure is a neural network, i.e. a mathematical model that tries to resemble the neural networks we have in our brains. In the case of generative AI, these networks can learn from a set of data and then generate new content.

One of the most effective models that allows this technology to develop content as images that appear authentic is that of Adverse Generative Networks, composed of a generator and a discriminator. The generator is the element that creates new content from a set of images/audio/texts/scripts that we provide. The discriminator on the other hand acts as a ‘judge’ and assesses whether the content produced by the generator is detailed and realistic enough.

It somehow trains and spurs it to do better. We have to imagine that this exchange of information between the generator and the discriminator goes on for a while and that the mechanism improves itself. This is how – for example – images become super-realistic.

Chat GPT is the most widely used generative artificial intelligence for text creation: it takes 72% of the market.

63% of users who use generative artificial intelligence to create images do so via Midjouney, a software that turns textual information into images. Some 15 billion images have been generated using text-to-image conversion algorithms. It took photographers 150 years, from the first photograph taken in 1826 until 1975, to reach the same threshold.

We have realized that generative AI is very powerful and is a great new step for content creation.

But are these works in any way protected by copyright?

Indeed, we have to remember that this technology is ‘trained’ by observing content that is already available online and that is generated by other users. This means that some of them are already copyrighted. But also that it was designed by a programmer who wrote the algorithm. And then there is also a user who solicits it by asking it to generate the content he wants. In short, the matter is not very simple.

And so, unlike traditional man-made works, the laws that exist are not always up-to-date to assign rights to those images. Who is the actual author of a machine-generated image? Let us say that in some cases, the law may consider the programmer as the author because he has provided the generative artificial intelligence with the instructions and parameters to generate the images or content. Then there is the aspect of the dataset with which these AIs are trained. These are often images or sounds that are themselves already copyrighted.

So is it legitimate to feed this content to an AI for it to train? Or is there a risk of creating a concatenation of copyright infringements?

Draft laws are trying to see whether new categories of copyright should be adopted for the works that are generated by these tools. The aim would be to enforce both the interests of those who develop AI, those who use it, and those of the artists from whom the AI takes information for training.

The speed with which generative AI is developing is astonishing: ChatGPT was released in November 2022, and only four months later, OpenAI released a new, large language model, GPT-4, with significantly improved functionality. The sectors in which generative AI is most used are marketing and advertising.

Generative AI can also be used to restore works of art and verify their authenticity through advanced image analysis and data comparison.

Imagine that you are an artist or illustrator creating images, paintings or photos that are entirely the result of your creative faculty. Then at some point, you find yourself with your work fed to a generative artificial intelligence that uses it to generate new content, some of it very similar to yours. How would you react? Would you feel robbed?

In November 2023, a group of artists filed a copyright lawsuit against Stability AI, Midjourney, and other companies, claiming that their AIs were misusing copyrighted work for training purposes. These illustrators claimed that the AIs of these companies create works using the same style and, more importantly, that those works are generated if the artist’s name is given as an indication. In short, even the connection to the authorship right of the work seems rather explicit.

In addition, the result is often that users end up generating works that would be ‘indistinguishable’ from their own. So much so that one of the artists allegedly found a work generated by Midjourney by searching the Internet for her name! However, once the case was heard, a US district judge dismissed some parts of the lawsuit but allowed the group of artists to pursue other complaints if they felt it was necessary.

The point is that these situations are often repeated. A group that manages writers also sued OpenAI, also in the US, on behalf of prominent authors, including even George R.R. Martin, the author of The Iron Throne. The accusation is that the company that owns ChatGPT trains its AI with the texts of these writers.

Nightshade is a new tool that allows artists to add invisible pixel changes to their work before uploading it online.

In this way, if inserted into an AI training set, it can cause the system to break down.

Nightshade exploits a security vulnerability in generative AI models since they are trained on large amounts of data, such as images retrieved from the Internet.

In 2021, regulation of artificial intelligence was proposed for the first time in the European Union.

After two years of discussion, on 9 December 2023, the Council and the European Parliament approved what has become known as the AI ACT, i.e. the world’s first legislation regulating the use of artificial intelligence in EU Member States. But beware: this does not mean that it is already reality!

After the agreement was made, experts and technicians from the institutions set to work to arrive at a final text that will be voted on by the European bodies.

So, theoretically, the legislation will come into force within a couple of years. But what does it envisage for generative AI and, above all, for copyright?

Generative AI, according to the European Parliament, should comply with transparency requirements.

This means that it is strictly necessary to disclose that the content has been generated by artificial intelligence design the model to prevent it from generating illegal content publishing summaries of copyrighted data used for training purposes. A crucial legal issue is precisely the use of copyright-protected works to train generative artificial intelligence models.

On the one hand, some think it is an infringement of copyright. On the other hand, some argue that it falls under one of the exceptions in the copyright directive. In particular, it would fall under the exception of ‘text and data mining’, i.e. the analysis of large amounts of content that allows researchers to discover patterns, trends, connections, and other in-depth information.

Currently, the ‘text and data mining‘ exception only applies in favor of research organizations and cultural heritage institutions, but there are experts in the field who would like to extend it.

In December 2023, the New York Times sued OpenAI and Microsoft for copyright infringement. The accusation is that they unlawfully used journalists’ content to train AI.

Imagine this scene: the most famous and, according to polls, also the most trustworthy newspaper in the world is suing the company holding the world’s most popular AI tool. On the one side, we have the New York Times, on the other OpenAI, the Chat-GPT company. In the middle is the Federal District Court in Manhattan, which has found itself taking on no small legal issue.

The problem, according to the NYT’s lawyers, is that copyrights were allegedly infringed on articles published by the newspaper. In particular, the AI would have been trained through the millions of texts written by the NYT journalists, which are of course copyrighted, without there being any agreement.

The New York Times is a newspaper with 170 years behind it and articles from 1996 to the present day are available online. All this material could potentially have been subjected to the artificial intelligence that allows us to use Chat-GPT, among other things with very good results.

But this bombshell did not explode out of thin air. The NYT states that months before the complaint, in April 2023, it had already contacted Microsoft and OpenAI to raise concerns about the use of its intellectual property. The idea was to try to find an amicable resolution together, reaching a commercial agreement and guidelines to be established for generative artificial intelligence. In the end, however, it seems that the talks were not successful.

OpenAI is valued by investors at more than USD 80 billion.

Microsoft has earmarked $13 billion for OpenAI and incorporated the company’s technology into its Bing search engine.

ChatGPT is a chatbot developed by US-based OpenAI in November 2022, the app was launched on iOS and Google Play Store in May 2023 and July 2023 respectively.

Between January and October 2023, mobile apps using the keywords ChatGPT and chatbot generated tens of millions of downloads.

The New York Times’ lawyers, in their complaint, claim that OpenAI and Microsoft not only use the journalists’ material without permission and free of charge but even use it to create products that can replace the newspaper, effectively taking away a slice of its audience.

Big things, in short. No specific monetary demand is included in the prosecution documents, but it is clearly stated that the defendants (i.e. OpenAI and Microsoft) should be held liable for ‘billions of dollars in legal and actual damages’. And there’s more: the explicit request to eliminate any chatbot model that uses copyrighted NYT material. Not least because the issue of the use of copyrighted content is compounded by another phenomenon to which regenerative AIs like Chat-GPT would be subject.

It is called ‘hallucinations‘, and is the mechanism by which chatbots respond to user requests with false information incorrectly attributed to a source. This would also be the case with the NYT, which would appear as a source on unverified content, often reporting false information that the newspaper did not cover. Sometimes AI algorithms produce outputs that are not based on the training data and are decoded incorrectly or do not follow any identifiable pattern. In other words, they ‘hallucinate’ and respond with false or invented information.

The New York Times has more than 11 million online and print subscribers. 60% of New York Times online readers are Gen Z and Millennials.

In early January, the official response from OpenAI arrived. Brace yourselves. According to the company, negotiations were underway for a partnership between the New York Times and OpenAI that would have given the newspaper, and I quote, ‘a new way to connect with its existing readership’.

On the other hand, Chat GPT users would have had access to reports written by the Times editorial staff. They also claim that in reality, the newspaper’s content did not contribute significantly to the formation of the company’s AI models. So, they conclude, the lawsuit they found themselves in would have been something of a nasty surprise. In short, they do not deny that some content was used but say that the use was partial. However, the accusations do not end there. According to OpenAI, there is a reason why articles from the NYT appeared in the answers provided by ChatGPT. The reason is that the NYTimes itself allegedly manipulated ChatGPT’s answers to make it respond by quoting material published in the newspaper itself.

The research centre that supports the work of the US Congress argued in a report that judgments in cases of intellectual property theft by AI could be the basis on which valid legislative action could be built in the US. There would be at least 12 subpoenas against OpenAI from June to December 2023.

The lawsuit between the New York Times and OpenAI-Microsoft will be one of the most closely watched in recent years. Commentators and analysts claim that it is the very first lawsuit by a true global giant of journalism against other giants, but of technology. The idea that an artificial intelligence could potentially exploit the NYT’s investment in its journalists for free and use it to create new products, without any payment or copyright protection, is a serious allegation.

Whatever the ruling will be, it will have a cascading impact on both Europe and our country, and – most importantly – it will set a compass for all the IP infringement trials and lawsuits that are already underway. Even if OpenAI invokes fair use, i.e. the American law that allows reworking of existing content (as long as it is not for profit), an ad hoc assessment will still have to be made.

It will have to be understood whether OpenAI has in some way gained from the possible use of the Times content and whether there has been at least an original reworking by AI or whether it is a copy, with some slight modifications.

One thing is certain. If the lawsuit had taken place in the European Union, where fair use does not apply, it would have been very difficult to prove that ChatGPT’s use of online content falls under one of the exceptions provided for by EU law.

Schickler works as a salesman in New York, but he comes into contact with Chat-GPT at a certain point. At that moment he realizes that he can use it to make one of his dreams come true: writing a book.

In 2023, nearly 1 in 4 U.S. writers used AI for their work.

Artificial intelligence tools have allowed us to create creative works such as books, producing entire parts of text thanks to a few simple instructions.

In just a few hours, Schickler managed to create an illustrated e-book for children. Thirty pages in total on sale on Amazon. Amazon allows self-publishing in a specific section of the site and therefore many are putting their books on sale there. Schickler’s e-book is called “The Wise Little Squirrel: A Story of Saving and Investing” and earned its author just under $100. It is a paltry sum, but – according to him – it was enough to push him to write again via AI. In fact, in an interview with Reuters, he said “I could see people building an entire career on this.”

As of February 2023 alone, Amazon had hundreds of ebooks for sale written by or via Chat-GPT. But the data is based only on those who said they use it.

21% of consumers are very concerned about the use of AI as a substitute for book authors.

An influx of suspicious AI-generated material has been put up for sale in recent months. In September 2023, Amazon found itself having to limit the amount of books self-published by authors who had used Chat-GPT.

The problem was the risk of incurring rights violations. Before imposing a limit of 3 books per author in one day, Amazon had already introduced the obligation for authors to inform the company if their content was generated by artificial intelligence. This is specified in the guidelines: it is essential to inform them if a new book or an update to an already published text has been written using AI.

In particular, “AI-generated” content is labeled differently from “AI-reviewed” content depending on the impact of the tool on the creation of the content. All these rules come after Amazon removed suspicious AI-generated books that were featured in a list of texts written by author Jane Friedman. Which is false.

A similar and certainly more dangerous thing happened with books, again written through AI, in which a fake mushroom expert suggested ways to collect them and even how to use them for psychedelic purposes. There was just a problem: the author, the images, and the descriptions were completely “invented” by artificial intelligence. Therefore they constituted a danger for anyone who might think that that advice, coming from an expert, was true and reliable.

Only 2% of Baby Boomers consider the use of AI in book writing positive, compared to 12% of Gen Z and 17% of Millennials.

69% of US writers believe AI is a threat to their work.

Among US publishing houses that had integrated AI tools into their systems in 2023, only 12% were using it for content production. Ideally, something is not created by a human being, it is simply not protected by copyright. Even if we take into account that there are differences between the US and European legal frameworks, the attribution to a human being remains valid.

From the Bern Convention for the Protection of Literary and Artistic Works to the 1996 WIPO Treaty, to the Trips, i.e. Trade-related Aspects of Intellectual Property Rights, it has never been thought that a work could be generated by a machine and protected by copyright, without there being a human author. It would therefore seem that works such as a book written by an AI do not enjoy copyright protection today. In particular, these works would not have those characteristics of creativity and originality that allow them to obtain protection.

But what about the person who provides input to the AI ​​to obtain the desired pieces of text? And how is the limit established between the requests made by humans and the machine’s ability to rework? We do not yet have clear answers to these questions.

Some experts argue that giving instructions to a machine is too little. The author would somehow appropriate the answer provided by the AI, claiming the right of paternity on this work without any proof that the same contents on which the AI ​​was based were free from copyright. Even if AI were based only on non-copyrighted content, for those who support this thesis, it would not be a human re-elaboration and therefore, no copyright.

29% of male and female consumers say they have no interest in reading a book written by AI.

Amazon is the ebook retailer most chosen by Italian consumers overall. We have seen how the law works today, but is it possible that the situation will be changed in the future? According to some experts, the answer is yes.

The risk of violating copyright by producing texts with AI is twofold. It exists both in the use of the data that is used to train the AI and in the use of the works generated by the AI. And if we add all this to the speed with which these tools are evolving, to their degree of customization, it will be fundamental to identify the cases in which there is a right to be protected and what type of rights we can talk about. Lawyer Daraffa, an expert on the topic, argues that if we reach the point where the production of content by AI will be recognized as an act of creativity, then it is likely that we will find ourselves having to talk about a new type of right connected to those works.

It would be, says the lawyer, something similar to the law that today protects the creation of simple databases, which protects the investment of energy and time that the producer of that set of data makes in the technology. However, if a right of this type is protected, there will be no copyright connected to the AI ​​itself.

From January 1, 2024, the first version of Mickey Mouse is no longer covered by copyright, so everyone can use it. But what is the public domain? Disney has 120 Mickey Mouse cartoons, with a 30-year gap between The Simple Things (1953) and Mickey’s Christmas Carol (1983).

Since January 1, 2024, after almost a century, the first version of Mickey Mouse has entered the public domain.

This means that it can now be adapted at will and reproduced without risk of legal action and without having to pay royalties. The first films in which the iconic Disney mouse appeared were called Steamboat Willie and Plane Crazy and were filmed in 1928. That is the beauty of 95 years ago. In fact, time is a crucial factor when it comes to copyright.

In practice, after a certain number of years from their creation, the works become free. In particular, 70 years must pass from the author’s death for works created in Europe and the United States. This means that all the works created in 1954 will become public domain this year. But be careful: not all versions of Mickey Mouse are in the public domain, only that of the very first two films, which first made Walt Disney’s great success. All the others are very different in appearance and characterization and therefore are the subject of different copyrights, still protected by law.

To become of public domain, 70 years must pass from the author’s death for works created in Europe and the USA.

So all the works created in 1954 in these territories will become public domain this year. Even the Peter Pan play, and its adaptations, also entered the public domain in January 2024.

The New York Public Library collects more than 180,000 pieces of public domain content online and is accessible to all.

An author can freely choose not to protect a work and to dedicate it to the public.

If there is no express authorization to place the work in the public domain, it can never be assumed that the work is free to use.

The public domain is the body of knowledge and information, such as books, images, and audiovisual works, which are not protected by copyright and which can be used without restrictions. The goal of the public domain is essential to the cultural memory and knowledge base of our societies.

It includes two categories of material.

The first is that of works for which copyright protection has expired. In most of Europe, copyright in a work lasts 70 years after the death of its longest-serving creator. Instead, if the copyright is held by a company, it lasts 70 years after publication. Once this temporary protection expires, all legal restrictions simply cease to exist.

The other category is made up of commons, essential information that is not covered by copyright. Let’s think about laws or judicial and administrative decisions: they are common goods that are too important for the functioning of our societies, therefore they remain in the public domain.

Beyond these cases, some several other limitations and exceptions reduce legal restrictions and ensure sufficient access to our shared knowledge and culture. The idea is to guarantee access and allow the functioning of essential social institutions.

For songs, copyright normally expires 70 years after the death of the song’s original artist.

Just because a song is in the public domain doesn’t mean every recording of that song is. If someone records a new version of a song that is in the public domain, that new recording of the song may still be protected by intellectual property.

The list of creative works whose rights expired in 2024 also includes The Circus, a film by Charlie Chaplin, and Orlando, a novel by Virginia Woolf.

Let’s go back to Mickey Mouse, in his 1928 version, and see what could happen to him. Over the years, Disney has been particularly careful to maintain the respectability of what is now its symbol. And he tried in every way to prevent the possible uses of this figure in inappropriate contexts and materials. However, when copyrights expire, things can change.

We already have examples of other fictional characters that have entered the public domain in recent years. In fact, in 2021 it was Winnie the Pooh’s turn. It took very little to transform the honey-hungry bear into something extremely different. It didn’t take long for someone to make Winnie the Pooh: Blood and Honey, a horror series where Winnie becomes a cannibalistic monster armed with a bat. And someone is already working on a film with the protagonist, Mickey Mouse himself, who however is… a serial killer. It’s called Mickey’s Mouse Trap, its trailer is already available online and the credits already read “This is not a Disney film or production. It is not affiliated or endorsed by Disney in any way.” In short: the producers of the film chose to avoid any problems with the house that first created the mouse.

Many public domain works can become ideas for derivative works, intellectual works created starting from one or more existing works.

The derivative work, if it comes from content not in the public domain, includes some aspects that can be protected by copyright.

West Side Story in its 1957, 1961, and 2021 adaptations, is considered a work derived from Shakespeare’s Romeo and Juliet tragedy. Many websites annually update the lists of contents that become public domain. And, although copyright protection is fundamental, when after many years a work becomes accessible and usable by everyone, many advantages are obtained. For example, screening rooms can be set up to watch films, orchestras, and bands can perform music publicly without paying licensing fees.

Even in online archives such as the Internet Archive or Google Books we can find entire works available online. And then many works that had fallen into oblivion due to their cost can also be recovered, to broaden their diffusion. If the main purpose of copyright is to promote and protect the creativity of individuals, the public domain plays a central role.

Copyright laws give authors important rights that encourage creativity and distribution, but on the other hand, they make these same rights limited in time. A limited time but long enough to allow the author to earn from his creations, and then to broaden the audience as much as possible when the rights to the work expire. It is a middle ground, which provides benefits on both sides.

Just think of the works of Greek mythology or those of Shakespeare which – being in the public domain – have inspired many adaptations, books, musicals, and films and are part of our widespread cultural heritage.

A reaction video is a video in which a creator films himself watching or listening to content created by others for the first time.

These videos include content belonging to other people, so copyright infringement is an imminent problem.

We all know what video reactions are, right? These are those formats, typically YouTube or Twitch, in which the creator reacts to the contents of other creators uploaded online.

Here you are. Have you ever wondered if they can use those images and videos? And how should a YouTuber behave if he sees that his content is being taken up by others?

Let’s start from the basics.

We have already seen that copyright protects those who hold the rights from both an economic and moral point of view. So every time a creator uses content produced by others and has not entered into agreements with these other people, he or she can potentially commit copyright infringement. However, when we talk about copyright on platforms like YouTube or Twitch, we often talk about copyright.

But copyright is something a little different from Anglo-Saxon copyright, although the two expressions are often used synonymously.

Law n°633/1941 in force in Italy guarantees the author of the work the ownership of both patrimonial rights and moral rights. Copyright, on the other hand, only concerns the protection of the economic rights of the author and is characterized as the exclusive right to copy the work. Furthermore, copyright is acquired only with the deposit of the work, a bit like what happens in Italy with the registration of a trademark or the patent request for an invention. Copyright, on the contrary, protects the work from its creation, regardless of whether it is later published.

Copyright is territorial: YouTube (or any platform) if it interacts with European users must guarantee copyright according to the legislation of the State in which the user operates, regardless of whether it then labels it as copyright.

To protect the copyright of third parties, according to YouTube policies, creators can only upload videos made personally or that they are authorized to use. Therefore, creators must not upload videos that they did not create themselves or use content in their videos whose copyright belongs to someone else, without the necessary authorizations.

During the first half of 2022 alone, YouTube received approximately 766 million copyright infringement reports.

Let’s see how it works for YouTube. The platform uses a series of copyright management and protection tools that allow rights holders to remain in control. These tools are varied. The first is the Content ID. The idea is this: when original content is created, the creator asks a YouTube manager to activate your content owner. This ID represents you in YouTube’s content management system. In this way, every time a creator uploads a new video the contents can be marked as copyrighted and inserted into the content management system. The content ID then analyzes user uploads and identifies if there are any matches. If matching videos are found, their ownership is automatically claimed. This means that the owner can choose whether to request to remove the “stolen” content or whether to impose monetization on that content and earn from it, precisely because the rights are his.

YouTube’s Copyright Match Tool also works similarly and automatically. However, this tool is activated only if there is an explicit request from a creator to remove content that they believe violates their copyright. But there’s more. There is the public DMCA (Digital Millennium Copyright Act) web form that is available to all 2 billion users who visit YouTube. If you use a copyrighted work in the content you share on Twitch, the person who owns the rights may send Twitch a copyright infringement notification, sometimes called a “takedown request” or “DMCA notice.”

Twitch users watched approximately 7 billion hours of streams in 2022.

As of October 2023, there were approximately 7 million streamers worldwide. On Twitch things aren’t too different. In fact, some streamers find themselves victims of “reaction thieves”, reaction thieves, who broadcast entire videos of others without making any significant contribution. While reactions can be fun and engaging, they should be a form of collaboration and mutual respect, not an excuse to use other people’s ideas and content.

Reactions to other people’s content are generally allowed but must be done in compliance with the rules.

First of all, it is advisable to contact the person responsible for the theft directly and ask for the material to be returned. If this does not lead to a resolution, you can use the complaint procedures by accessing the “copyright” tab of your account. Twitch will process the complaint and take necessary action, which may include removing the content in question or even deleting the account of the user responsible for the stolen content.

Twitch also uses the DMCA. If a creator thinks their work has been used without permission, you can file a DMCA complaint. Although the DMCA is a United States law, many countries have laws that work similarly. So sometimes, for simplicity’s sake, Twitch uses terms like “DMCA” and “DMCA Notice” to apply to all accounts. This process involves filing a formal request to remove the content to which the infringing user can respond by filing a counter-notice.

An artist can rework and artistically interpret original works and the creator of the derivative work can acquire copyright on its elaboration. However, the rights owned by the author of the original work must be respected, and this processing cannot be done without his consent.

Some national courts and the Court of Strasbourg have confirmed that the prior consent of the author of the original work is not necessary if the appropriation of the latter occurs to distort its meaning for caricatural, parodic, or satire purposes.

There are exceptions to copyright infringement, especially in the case of reactions. Article 70 of the Italian copyright law clearly states that there are circumstances in which there is no violation of copyright. The summary, quotation, or reproduction of passages or parts of the work and their communication to the public are free if carried out to criticize or generate discussion.

Provided that this activity is carried out within the limits justified by these purposes and provided that they do not constitute competition with the economic use of the work. There is no violation if the use of content is carried out for teaching or scientific research purposes; the use must also take place for illustrative purposes and for non-commercial purposes. Free publication via the internet, free of charge, of low resolution or degraded images and music, for educational or scientific use and only if such use is not for profit.

The summary, quotation, or reproduction of content must always be accompanied by a mention of the title of the work, and the names of the author and the publisher. And if it is a translation, the name of the translator must also be entered, if the indication is present on the reproduced work.