OpenAI, known for its ChatGPT models, faced scrutiny as sources revealed the Defense Department’s testing of Microsoft’s version of OpenAI technology, despite OpenAI’s ban on military use. The controversy arose after OpenAI’s deal with the US military, prompting internal criticism and calls for transparency from CEO Sam Altman. In 2023, OpenAI explicitly prohibited military access to its AI models, yet the Pentagon had already begun utilizing Azure OpenAI, a Microsoft-offered variant of OpenAI’s technology. This revelation raised questions about the clarity of OpenAI’s usage policies and the involvement of Microsoft, the startup’s major investor with licensing rights.
While some OpenAI employees expressed wariness towards Pentagon ties, confusion prevailed regarding the applicability of OpenAI’s policies to Microsoft’s products. OpenAI and Microsoft clarified that Azure OpenAI products were not bound by OpenAI’s restrictions. Microsoft’s spokesperson emphasized that the Azure OpenAI Service, available to the US Government since 2023, operated under Microsoft’s terms of service. Notably, Microsoft refrained from specifying when the service was accessible to the Pentagon, highlighting that it did not have ‘top secret’ approval.
Source: WIRED