Staffers of the European Commission were advised by internal memo to consider whether generative AI might be violating intellectual property rights before using its output in official communications because protected content may be used to train the model. Also, outputs may not credit protected material, so users may not know to obtain permission for use from rights holders.
This as part of the Commission’s guidelines issued on May 30 “to help staff members assess the risks and limitations of online available generative Artificial Intelligence (AI) tools and set conditions for their safe use in working activities of the Commission.”
While the EC notes how generative AI tools can boost effficiency and productivity, they also carry risks and limitations, according to reports by parties who have seen the 4-page document.
One risk is the disclosure of sensitive information or personal data that may be generated and subsequently used to feed future generated outputs that may become public. Therefore, staff can’t use “any information that is not already in the public domain, nor personal data, with an online available generative AI model.”
Another risk is that due to potential shortcomings of AI models, the AI’s responses may not be accurate, and may carry biases.
The guidelines urge workers to apply their own critical thinking skills to evaluate the output before using it. Workers should never “cut and paste” AI-generated output into official documents, and should avoid using AI tools when working on “critical and time-sensitive processes.”
The guidelines also issue an exception for their own AI services, saying that are “not necessarily relevant for internally developed generative AI tools from the Commission. Internal tools developed and/or controlled by the Commission will be assessed case by case under the existing corporate governance for IT systems.”
EU Commission issues internal guidelines on ChatGPT, generative AI. Article. May 31, 2023. Luca Bertuzzi. Euractiv
European Commission tells staff not to use generative AI. Article. June 1, 2023. Juseph Brunoli. Techzine
Why it matters
A note issued alongside the guidelines say that “The guidelines … aim at assisting European Commission staff in understanding the risks and limitations that online available tools can bring and support in appropriate usage of these tools.
The guidelines were adopted by the EC’s Information Management Steering Board at the end of April and are positioned as a ‘living document.’