MOW’s legal counsel, Preiskel & Co, draw our attention to a recent opinion from EU Advocate General Szpunar which has placed the weight of EU legal opinion behind MOW’s Search Only Contract (SOC) and confirms that AI platforms can breach copyright.
The SOC, which allows publishers to quickly and easily charge AI companies for content harvesting, is a machine readable contract that is put in place via a website’s Robots.txt file and Terms and Conditions pages. The new opinion supports this exact means of informing bots of a publisher’s intentions about AI harvesting. The opinion states that any opt-out mechanism should ‘not be excessively complicated’ and should be ‘machine readable’. Furthermore, AI bots can be informed of a site’s decision to opt out through the ‘Robot Exclusion Protocol’ (Robots.txt) or in a website’s terms and conditions. The opinion also supports the idea that an opt-out can be expressed in natural language ‘given the current natural language processing capabilities of LLMs’.
This legal support for the SOC’s implementation is significant for two reasons. Firstly, publishers can now roll out the SOC confident in the knowledge that Europe’s most senior legal figures support its approach and, secondly, that the AI companies will not be able to delay any proceedings that arise from the SOC by bringing into question the logistics of its implementation.
Beyond the SOC, the opinion also raises significant issues for AI providers by finding that AI services likely violates news publishers’ EU copyright law rights when they reuse an article’s distinctive expression so extensively that its answer substitutes for consumers reading the original. This fundamentally underpins the rights of publishers and raises significant questions for AI companies who claim that their harvesting and use of others’ content is somehow exempt from established copyright norms.
For more detail, read Preiskel’s blog here.