The Swedish government has banned the use of the "ChatGPT" and "Claude" artificial intelligence models in several government agencies due to security concerns and the potential for confidential information leaks. The restriction went into effect on Monday.
According to Swedish state radio, employees in these agencies have been advised to use Microsoft's "Copilot" AI model instead of the two banned applications. The decision follows growing concerns about how government employees are using generative AI tools.
Swedish officials are worried that employees might unknowingly feed confidential and sensitive information into applications whose parent companies do not have specific data protection agreements with the Swedish government. This could jeopardize national security and citizen privacy.
These concerns escalated after an internal review at the Swedish Transport Agency revealed that its employees had been using about ten different AI tools. Investigators found that in some cases, sensitive government information was entered into these programs, and the AI-generated responses were not used correctly or with the necessary diligence.
Mikael Wetercrantz, director of digitalization for Sweden's government offices, warned that using AI tools without a formal agreement with the provider companies could pose serious challenges to the government's information security.
In contrast, the Swedish government has authorized the use of Microsoft's Copilot application. This program was approved after thorough security reviews, and additional measures have been implemented to protect the information entered into it.
The ban only applies to a number of Swedish government agencies, and ordinary citizens can still access ChatGPT, Claude, and other AI tools for personal use.
Source: Swedish state radio
🌐
