The West is not a bloc when it comes to data. The EU and US have fundamentally different views on the use of data, and this is incompatible with the way data is processed today. Where does this difference come from and what is the solution?

ChatGPT exists because of everything you and the rest of the world have put on the Internet. The data set that OpenAI uses Great language model (LLM) trained and was removed from the network. They haven’t given permission for this and it’s actually not allowed in the EU. But OpenAI has usurped the Internet from the US, where this is not a problem.

Is privacy dead?

This is just one example of the contradictions between the USA and the EU. Across the ocean, they have no problem with European data going directly from the transatlantic fiber optic lines into the NSA databases. “Privacy is dead,” Meta CEO Mark Zuckerberg previously said. Earlier this year, at an event hosted by data cloud specialist Snowflake, we got into a discussion with Americans about “the EU’s annoying inhibitions on data collection.” Europe protects its data, especially that of individuals, while the US pursues a more liberal strategy.

On the sidelines of Isaca’s Digital Trust World conference in Dublin, we speak to Dr. Valerie Lyons, Director of Operations at BH Consulting and cybersecurity and data protection specialist.

Retroactive restoration

“It actually happened and there’s nothing we can do to change it,” confirms Lyons when we talk about OpenAI’s scraping work. “Right now we are trying to retroactively solve a problem while the system is already broken. Data scraping There is no legal basis in the EU, but once data has been used to train a model, you can no longer use it.”

Right now we’re trying to retroactively fix an issue while the system is already broken.

Dr. Valerie Lyons, Operations Manager BH Consulting

The EU is currently working on a law to clarify the situation. “A complete ban on scratching is not an option. It’s allowed in the US, so it will happen there too. If we ban it here, we will victimize our own companies,” Lyons said. Balancing privacy, functionality and innovation will be a delicate task.

Difference in definitions

Part of the problem lies in a fundamental difference in what data protection and privacy actually mean. “When we talk about data protection, we actually mean everything that has to do with data and privacy. Data protection itself is a much broader term that is anchored, among other things, in the European Convention on Human Rights. In the USA, data protection has very specifically to do with control mechanisms. “We talk about data protection as an independent topic.”

“The use of data is a fundamental problem that we need to solve,” she continues. And to understand this problem, we need to look at the origins of the different views on data and its privacy. Lyons delves into history.

Data, Nazis and terrorists

“Our understanding of data and privacy comes from the post-Holocaust era,” explains Lyons. “Among other things, the misuse of personal data in the population registers of villages and cities has a significant impact. Combating the misuse of personal data remains a priority in the EU today. That’s the priority.”

When we cross the ocean something different happens. Lyons: “There was no Holocaust in the USA, but there were terrorist attacks.” Among other things, the attacks of September 11th shaped America’s view of privacy. “In general, Americans are much more accepting of an invasion of their privacy in the name of public safety. They want to feel protected.”

There was no Holocaust in the United States, but there were terrorist attacks.

Dr. Valerie Lyons, Operations Manager BH Consulting

Because we are afraid of abuse, they ask for protection. Both camps value privacy, but approach the issue from a completely different perspective. Lyons is aware that this vision represents the bigger picture, although there are of course other visions here and in the United States. “Age and political beliefs play an important role,” she adds.

Check

The solution will probably be a compromise. “We need to ensure that people are rewarded for the data they share.” This is only possible if controls are put in place to effectively determine what you want to share and what you don’t. A practical way to achieve this for data scraping is to add some type of cookie to a website that allows or disallows this, analogous to robot.txt This already tells bots whether a page can be indexed or not.

Solutions are in the works, but it is unclear how far they will go. The fact is that there is a difference in perspective that can be explained by the history of both regions. When it comes to something sensitive like personal data, it’s always good for both parties to know the underlying motivation.