Tech Firms Support European Commission's Artificial Intelligence Ethics Guidelines by Endorsing the Code
In a recent development, Google's parent company Alphabet has expressed concerns about the potential impact of the European Union's AI rules on innovation. Kent Walker, the president of global affairs at Alphabet, voiced these concerns in a blogpost on Wednesday.
The European Commission released a voluntary AI Code of Practice on General-Purpose Artificial Intelligence (GPAI) earlier this month. Companies joining the Code of Practice from 2nd August 2025 need to comply with specific AI Act obligations mainly focused on transparency, copyright, and for some, safety and security.
Under the transparency requirements, providers must maintain detailed technical documentation, provide comprehensive information to downstream deployers, and publish summaries of training data. This transparency aligns with Article 53(1) of the AI Act and ensures users and deployers understand the AI model's capabilities and limitations.
In terms of copyright compliance, providers must respect restrictions on data access used for training, such as paywalls or subscription models. They must also implement technical safeguards to prevent outputs reproducing copyrighted content protected by EU law, and draw up and enforce a copyright policy compliant with EU legislation.
For models classified as having systemic risk, additional voluntary obligations related to safety and security measures to mitigate such risks are required.
The Code of Practice is voluntary but provides a structured and recognized way for companies to demonstrate adherence to AI Act obligations. It is expected to be formally adopted by the European Commission soon after its publication. Signatories include major AI providers like Anthropic, OpenAI, Google, Amazon, and IBM.
However, Meta (previously Facebook) has announced that it will not sign up to the Code, citing restriction of innovation. Despite this, Meta will still need to comply with AI Act obligations that start applying on 2nd August.
The AI Act obligations concerning transparency and copyright will need to be demonstrated by xAI, as it has only signed up to the Code's Safety and Security Chapter. Alternative means will be required by xAI to demonstrate compliance with the AI Act's obligations concerning transparency and copyright.
The European Union member states should have appointed national oversight authorities to ensure businesses comply with the AI Act from Saturday, 27th of this month. National oversight authorities will be responsible for ensuring businesses in their respective countries comply with the AI Act.
The final version of the AI Code of Practice on GPAI comes closer to supporting Europe's innovation and economic goals, according to Google. Companies must ensure robust transparency documentation, enforce copyright protection policies, and, where applicable, implement safety and security measures to comply with the AI Act obligations stipulated in the GPAI Code of Practice starting 2nd August 2025. Other providers can sign up to the AI Code of Practice on GPAI later.
- Google, being a significant player in the artificial-intelligence sector, must ensure robust transparency documentation, enforce copyright protection policies, and implement safety and security measures to comply with the AI Act obligations stipulated in the GPAI Code of Practice, starting from 2nd August 2025, as stated by Kent Walker, the president of global affairs at Alphabet.
- The European Union's AI rules, which emphasize transparency, copyright, and safety and security, could potentially impact the advancement of technology, particularly artificial-intelligence, as expressed by Alphabet, Google's parent company, due to concerns about innovation.
