On Wednesday a German court issued a preliminary ruling against Google, finding the company liable for false statements generated by its AI overview search feature. While the court correctly suggests that AI is not a necessity for searching the web, the question then is of its added value.
Denas Grybauskas, Chief Governance and Strategy Officer at Oxylabs, web intelligence company, shares his insights on AI impact in search and a critical tension between rapid technological deployment and the fundamental utility of the internet.
The court’s ruling suggests that “nobody needs AI to search the internet.” – Would you agree with this claim?
The ruling correctly identifies that “search” as we know it—finding a specific URL—doesn’t fundamentally require AI. However, the modern user isn’t just looking for a link; they are looking for synthesis and answers. For AI to provide those answers reliably without hallucinations, it must have an uninterrupted pipeline to the open internet. If we decouple AI from the public web, we risk creating models that operate in a vacuum, relying on static, outdated training sets that inevitably lead to factual errors.
What are the broader implications of this ruling for AI companies building the next generation of conversational and search tools?
If the ruling is enforced through the legal system, it will expose AI companies to liability for AI speech. This decision places a higher premium on AI safety and the reliability of the underlying information ecosystem. Moving forward, AI developers will need to prove their systems are not just “smart,” but fundamentally safe and verifiable. This means moving toward rigorous, controlled data curation that minimizes hallucinations and mitigates the risk of propagating misinformation, as the legal consequences for AI speech are only going to grow more significant.
If courts continue to push back against AI integration in search, what is the most reliable way for developers to ensure their models stay accurate?
Developers must prioritise transparency and diverse sourcing. Rather than relying on a handful of high-profile data deals—which can lead to a “point of view” bias—they should use multiple sources to build models that verify all strong statements against a variety of perspectives, prioritising the most reputable sources. AI answers might become less confident in their tone, but more reliable and nuanced.
About the expert:
Denas Grybauskas is Oxylabs’ Chief Governance and Strategy Officer, leading legal, risk management, ESG, and communication teams. Denas is also a global thought leader, providing commentaries to the media, and an educator, sharing his knowledge with students and professors at numerous prestigious universities, such as the University of Michigan. Additionally, he is a major voice of the Ethical Web Data Collection Initiative (EWDCI).
Expert on German court vs Google: modern users need more than just URLs
Posted in Commentary with tags Oxylabs on June 12, 2026 by itnerdOn Wednesday a German court issued a preliminary ruling against Google, finding the company liable for false statements generated by its AI overview search feature. While the court correctly suggests that AI is not a necessity for searching the web, the question then is of its added value.
Denas Grybauskas, Chief Governance and Strategy Officer at Oxylabs, web intelligence company, shares his insights on AI impact in search and a critical tension between rapid technological deployment and the fundamental utility of the internet.
The court’s ruling suggests that “nobody needs AI to search the internet.” – Would you agree with this claim?
The ruling correctly identifies that “search” as we know it—finding a specific URL—doesn’t fundamentally require AI. However, the modern user isn’t just looking for a link; they are looking for synthesis and answers. For AI to provide those answers reliably without hallucinations, it must have an uninterrupted pipeline to the open internet. If we decouple AI from the public web, we risk creating models that operate in a vacuum, relying on static, outdated training sets that inevitably lead to factual errors.
What are the broader implications of this ruling for AI companies building the next generation of conversational and search tools?
If the ruling is enforced through the legal system, it will expose AI companies to liability for AI speech. This decision places a higher premium on AI safety and the reliability of the underlying information ecosystem. Moving forward, AI developers will need to prove their systems are not just “smart,” but fundamentally safe and verifiable. This means moving toward rigorous, controlled data curation that minimizes hallucinations and mitigates the risk of propagating misinformation, as the legal consequences for AI speech are only going to grow more significant.
If courts continue to push back against AI integration in search, what is the most reliable way for developers to ensure their models stay accurate?
Developers must prioritise transparency and diverse sourcing. Rather than relying on a handful of high-profile data deals—which can lead to a “point of view” bias—they should use multiple sources to build models that verify all strong statements against a variety of perspectives, prioritising the most reputable sources. AI answers might become less confident in their tone, but more reliable and nuanced.
About the expert:
Denas Grybauskas is Oxylabs’ Chief Governance and Strategy Officer, leading legal, risk management, ESG, and communication teams. Denas is also a global thought leader, providing commentaries to the media, and an educator, sharing his knowledge with students and professors at numerous prestigious universities, such as the University of Michigan. Additionally, he is a major voice of the Ethical Web Data Collection Initiative (EWDCI).
Leave a comment »