AI Act draft passed in EU, generative AI copyright in the spotlight
posted Tuesday May 2, 2023 by Scott Ertz
As the interest in and usage of artificial intelligence (AI) has increased, so has concern over its safety and future. People ranging from Elon Musk to the engineers at Google have been speaking out over fears about what could happen if AI goes unchecked. Some governments have just begun to think about the future of AI in the world, but the European Union was ahead of the curve and has passed a draft version of the Artificial Intelligence (AI) Act.
Fears over AI
Fears over AI have been growing in recent months. In March, a letter from many prominent people in the technology world was published, encouraging a pause in AI development. The letter was met with mixed results. Some agreed with the letter as written. Others thought 6 months wasn't long enough. Still, others thought it was impossible to put the genie back in the bottle, so we should just keep going, as-is.
Going further, some have spoken out in great detail about the dangers of AI. Last year, before any of the hype around the technology had really gotten moving, an engineer at Google spoke out saying that the Google AI chat technology had become sentient. His arguments were dicey, especially knowing what we know about how the technology works. But, it's a consistent fear with AI, being a long-time sci-fi trope: AI becoming sentient and either dealing with its own death or dealing death in destruction of its own.
Also coming from Google, Geoffrey Hinton, known as the Godfather of AI, has left the company. He has spent the last 5 decades developing and cultivating AI technologies, like what underpins our current trend of generative AI. However, he has now left Google citing fears over how the technology can cause serious harm to people and humanity itself.
The EU to the rescue?
The European Union has passed a draft version of the Artificial Intelligence (AI) Act. This act is intended to protect people, in various ways, from the encroaching technology. It includes GDPR-style protections, as well as disclosures when a model has used copyrighted materials.
The draft is just the latest step in the process. Now, the member states will need to weigh in on the proposal. Each state will be able to pitch amendments to the draft, adding or removing considerations. Then, the EU Parliament will vote on the finalized version of the bill, which would then be implemented across Europe.
Since the draft, many organizations have pitched the issues they feel are not being adequately addressed. Publishers, like ourselves, think the provision about disclosure is not enough. Rather than just disclosing to the public when copyrighted content has been used without permission, the organizations should be required to obtain permission to index that content in their models.
Amnesty International has also weighed in, saying in a blog post,
The AI Act offers EU lawmakers an opportunity to put an end to the use of discriminatory and rights-violating artificial intelligence (AI) systems.
As this is a big and important topic for many, this is not likely to be a short process. For reference, it took around 6 months for all of the member states to weigh in on the proposal to require USB-C as the universal charging solution in the EU. So, expect this to be a long process, but we will be following it closely.
