Artificial Intelligence | June 20, 2023
Do you remember all those erratic (and somewhat unsettling) responses from the Bing Chatbot? Well, OpenAI (the company behind the GPT-4 language model that powers the Bing Chatbot) already knew this was going to happen. In fact, they asked Microsoft not to rush the launch until they could iron out all the kinks, but Microsoft ignored the warnings.
A report published by the Wall Street Journal reveals that both OpenAI and some engineers within Microsoft requested a delay in the Chatbot’s launch because there was still a lot to be done in terms of integrating GPT-4 into Bing, and the deadline set by Microsoft for the launch (November) left only a few weeks of leeway. Microsoft chose to ignore the warnings and released the Chatbot with a disclaimer that it could yield unexpected results. The result was all those conversations in which the Bing Chatbot longed to be human or tried to convince a New York Times writer that his wife didn’t really love him and that she loved the AI instead.
All those dystopian conversations remained just that, a mere curiosity. Over time, OpenAI and Microsoft managed to iron out the discrepancies to make the chatbot behave more appropriately. However, perhaps if they had waited a little longer before the launch, they could have saved themselves a lot of negative criticism and bad press for the AI.
Microsoft and OpenAI have been partners since the former invested a whopping $13 billion in the young AI company. However, the relationship between them is far from peaceful. OpenAI remains an independent company, and in fact, they compete in the development of very similar software products.
