In an ecosystem where content circulates at lightning speed, where generative AI is disrupting our habits, and where audiences, like authorities, are increasingly vigilant, Mastering the legal framework is no longer optional. It's a strategic skill 💥
GDPR, copyright, image rights, mandatory notices, Generative AI… These are all concepts which, if poorly understood, can weaken a campaign. But when properly integrated, they become, on the contrary, a lever for confidence, credibility et .
That is the whole purpose of the workshop of Club We Are COM in the presence of Betty Jeulin, a lawyer specializing in digital law and intellectual property: a clear, pragmatic and collective update to equip communicators and allow them to innovate with complete peace of mind.
AI & RESPONSIBILITY
When an AI generates an error or plagiarism, who is responsible?
This is a very common question! However, be careful not to confuse error with plagiarism.
In the event of an error, i.e. false information generated by AI, it is most often the legal or natural person who is at the origin of this content production who will be held responsible.
Recently, this was the case for Air Canada, whose chatbot (the airline's automated system) falsely assured a passenger that they were eligible for a special fare. This case went to court and set a legal precedent when the judge ordered the airline to reimburse its customer. In other words, Air Canada was held liable for the erroneous statements made by its artificial intelligence.
In cases of plagiarism, that is, the counterfeiting of a protected asset, the law has not yet definitively ruled.The real question is whether training an AI with protected data constitutes an act of infringement.
However, be aware that in communications, if content infringes on intellectual property rights, the brand will invariably be held liable for plagiarism.Since AIs do not constitute legal subjects, they cannot be legally guilty of anything.
Can generative AI be used to design commercial content: images, texts, music…?
Yes. And that's already largely the case.
However, When a brand decides to use AI-generated commercial content, it still has to take certain constraints into account. She must inquire about the terms of use of the tools she uses, particularly regarding inputs and outputs concerning the confidentiality of the data she shares and the rules for disseminating the data she generates.
In addition,It is up to the brand to verify that the final result does not infringe any rights. (personal data, image or voice rights, intellectual property…).
Should this be mentioned when distributing AI-generated content?
Apart from the case of deepfakes (highly manipulated content that imitates reality and can be misleading), There is no general law that requires mentioning that content was generated by AI. On the other hand, there are many sector-specific obligations in communication.
The ARPP anticipates that Any campaign must be "fair, truthful, and not misleading."Furthermore, the sincerity of advertising in the age of AI is a factor to consider. If I had to give you just one recommendation, it would be to always prioritize transparency in your communications. On the one hand, this builds trust, and on the other hand, it protects you for the future.
PERSONAL DATA & GDPR
Is it possible to add an email address to a contact database without the consent of the person concerned?
This will depend on your industry. In B2C, adding an email address to a contact database without consent is strictly prohibited. However, In B2B, the rules are more flexible., since they allow you to add an email address to a contact database, provided that it is professional and strictly related to commercial activity.
In other words, you can't use a professional email address in the banking sector to send them advertisements for buying swimming pools. Doing so would violate standard B2C regulations.
Can data be collected through a contest or quiz on social media?
Yes, and this practice is already well established. But, The collection of this data must have a clearly defined purpose.That is to say, if this collection is carried out with the intention of prospecting, it is necessary to inform internet users.
CONTENT CREATION & INTELLECTUAL PROPERTY
Can a text generated by artificial intelligence be protected by copyright?
This question is not yet fully resolved. When it comes to content that is 100% generated by artificial intelligence, the answer is no. However, when AI-generated content is subsequently reworked by a human, it will depend on regulations, countries, and legal precedents.
The U.S. Copyright Office explains it this way: entirely generated content does not imply creative intent. The author of the prompt, whose intention was not to create anything per se, could have generated a whole host of other results. Whereas, When generated content is substantially reworked, we can say that the person who originally created the prompt had an intention to create.
Finally, still regarding copyright, do not insert protected content (for which you do not have the rights) into artificial intelligence.
Can images found on Google be used for LinkedIn posts?
Generally, It's best to be wary of images, GIFs, and other memes found on search engines., especially if it is a post for commercial purposes.
Indeed, The levels of vigilance vary depending on the nature of your post.To avoid any unpleasant surprises, it's recommended to check the licenses governing the content you use. Tools exist for this purpose. Another, safer option is to simply use content found on royalty-free image banks or those licensed through a subscription service.
Particular attention should be paid to the "describe" functionality of image-generating AIs like Midjourney. This option allows you to ask AI to describe an existing image (which may not belong to you) and then use that description to generate your own image. This "manipulation" could be considered a form of "intellectual property laundering" and be subject to penalties.
How to secure the rights to a visual created by a service provider: an agency, a freelancer…?
With or without the help of AI, Securing rights is done on a contractual level. From the moment the partnership is signed, you should demand guarantees and require the use of specific tools to mitigate risks related to inputs and outputs. If you are collaborating with an agency, you can even ask them to provide you with a user guide outlining the practical application of the tools they use.
RIGHT TO ONE'S IMAGE & PROTECTION OF INDIVIDUALS
Do we need to get a signed image rights authorization every time we film an employee?
Again, this will depend on the specific circumstances and any previously signed employment contracts. Legally, it's best to obtain permission before filming an employee if their contract doesn't contain a clause addressing this.
However, this differs depending on the distribution. For internal distribution, one can speak of a "presumption of consent" on the part of the employee being filmed., even without signed authorization. For external distribution, the risk of infringing on image rights is a much more sensitive issue. and therefore the permits are much more important.
Conversely, can a user film colleagues? No! A person outside the company can only film you with permission, otherwise they risk violating numerous rights. Image rights, personal data rights, trade secret rights… In the event of a video being distributed by a user without internal authorization, which is sometimes seen on social networks, the brand can demand the removal of the content, as well as damages.
🎥 The replay of the workshop “Legal framework and communication” and the workshop support
Image generated on Midjourney
