At the beginning of this year, OpenAI, the company behind ChatGPT, announced the new usage policies for their generative AI, highlighting those cases in which individuals involved in the 2024 United States elections would not be able to use them. One of them specified that no political campaign or lobbying group could use their services.
But a candidate for the US presidency would have ignored these regulations and tried to bypass them, creating a chatbot that could have used the OpenAI ChatGPT model through third parties.
According to Wired, the presidential campaign of Robert F. Kennedy Jr. through a third-party candidacy created a chatbot to help explain his views, using Microsoft’s Azure OpenAI Services to access a third-party chatbot called LiveChatAI.
In a statement, a spokesperson for the Kennedy campaign said that the chatbot was used as an “interactive FAQ for our followers” and that they believed it was “a great help in getting the information they need on the go”.
Although neither Azure OpenAI Services nor LiveChatAI have policies that prevent them from being used by electoral campaigns, the LiveChatAI website itself states that it is “powered by OpenAI technology“.
In an email, the co-founder of LiveChatAI, Emre Elbeyoglu, stated that their service also uses other large language models like Llama (from Meta) and Mistral, and added that the company “could neither confirm nor deny” any information about the use of their services by their clients due to “our commitment to client confidentiality”.
Microsoft also confirmed in a statement that the use of Azure OpenAI Services by Kennedy’s presidential campaign (descendant of the famous former president) did not violate its policies and was not subject to OpenAI’s regulations on the use of its generative AI tools. However, Wired reports that the chatbot has been removed from Robert F. Kennedy Jr.’s official campaign website and can no longer be used by anyone.
