[DSC Europe 23][Pandora] Nikola_Kliska_Andrea_Radonjanin_AI_-_Legal_Insights.pptx
1. AI – Legal Insights
Andrea Radonjanin, Head of IP practice, Partner at Moravcevic
Vojnovic and partners in cooperation with Schoenherr
Nikola Kliska, Head of IP, Karanovic & Partners
2. 2
Copyright Issues in Input Data for Generative AI
o Data Ownership
o Fair Use and Transformative Use
(US/EU/Balkan approach)
o Court cases
o Data Licensing
o Public Domain Data
?
3. 3
Types of Protection for AI Systems:
Databases
Trademarks
Patents
Copyright Protection
Trade Secrets
5. 5
Copyright Issues in Output Data for Generative
AI
o Output Ownership
o Status of Output works
o Infringement of earlier works by Output works
6. 6
AI itself?
Authors whose works were
used for training?
AI developer?
1. Who owns the Output?
User/prompter?
7. 7
How much human authorship is required for a user of an AI
system to qualify as the owner of AI-generated output?
How far from the creative input can the human sit to maintain
authorship?
Is generative AI truly
creative?
2. Output status – what is created?
9. 9
Does Output copy or resemble earlier works?
Does Output “in style of”
infringe?
Does Output imitate earlier works?
3. Does Output infringe upon earlier works?
10. 10
Practical considerations
• Humans are indispensable
• Examine Terms of Use
• Consider jurisdictions in which AI-assisted
works may be copyrightable
• Alternative protections (trademark, trade
secret and contractual protection)