Great read. What about who owns the output from General AI systems? If a developer uses GIT AI to generate code (data). Couldn't Git argue that because the dev used their AI models that were generated on data they owned. Doesn't that mean the output is now owned by Git?
Also who owns the copyright to photos that have been modified by AI?
These are all great questions for which we don't yet have solid answers. I think it is not only about ownership but also, for example, about responsibility for data quality.
It's the AI vendor responsibility to ensure the best models. It's the customers responsibility to ensure the output meets or exceeds their use case. Unless the AI vendor is guaranteeing the output for quality. Which they won't.
great points ! I think the key issue is that AI needs data and lots of it to function effectively and this is going to need a relook at how we manage data security. Attacks like data poisoning are only going to increase going forward as well as the privacy implications of storing so much information
Data-based attacks will also become much more sophisticated. We will need new tools and practices to effectively support the security of data supply chains.
Great read. What about who owns the output from General AI systems? If a developer uses GIT AI to generate code (data). Couldn't Git argue that because the dev used their AI models that were generated on data they owned. Doesn't that mean the output is now owned by Git?
Also who owns the copyright to photos that have been modified by AI?
Lots of questions around data and ownership.
These are all great questions for which we don't yet have solid answers. I think it is not only about ownership but also, for example, about responsibility for data quality.
It's the AI vendor responsibility to ensure the best models. It's the customers responsibility to ensure the output meets or exceeds their use case. Unless the AI vendor is guaranteeing the output for quality. Which they won't.
great points ! I think the key issue is that AI needs data and lots of it to function effectively and this is going to need a relook at how we manage data security. Attacks like data poisoning are only going to increase going forward as well as the privacy implications of storing so much information
Data-based attacks will also become much more sophisticated. We will need new tools and practices to effectively support the security of data supply chains.