A year ago, ChatGPT did not exist to the general public. Today, it is one of the most talked-about artificial intelligence products in the world, and the person responsible for it is a 34-year-old engineer Mira Murati. Ms Murati was today appointed as the interim CEO of OpenAI after the board lost confidence in Sam Altman.
Not only ChatGPT, Ms Murati was also responsible for promoting Dall-E, an AI model that generates images from text. Both the offerings by OpenAI have come into the limelight after deepfake videos of actors Rashmika Mandanna, Katrina Kaif, and Kajol went viral.
Prime Minister Narendra Modi also expressed concern yesterday over the misuse of AI and urged the ChatGPT team to put riders in place to stop the creation of such morphed media.
Mira Murati, speaking on a talk show, explained how the company is working to prevent the creation of morphed media using their technologies.
“We have chosen to make Dall-E available to the public but with certain guardrails and with certain constraints,” she said while speaking to comedian Trevor Noah last year.
“We do want people to understand what AI is capable of. But right now, we don’t feel very comfortable around the mitigation of misinformation, and so we do have certain guardrails,” Mr Murati added.
The 34-year-old explained how they regularly eliminate certain data to ensure that users cannot generate images of public figures.
“We do not allow generation of public figures. So we will go into the data set and eliminate certain data. That’s the first step – looking at the training data of the model and auditing it, making interventions to avoid certain outcomes,” she said.
“Later, we will look at applying filters, so that when you put in a prompt, it won’t generate things that contain violence or hate,” Mira Murati added.
A series of deepfake videos on social media with morphed faces of actors Rashmika Mandanna, Katrina Kaif, and Kajol have sparked concern about the misuse of AI. Several voices in the film industry, including legendary actor Amitabh Bachchan, have called for legal action.
The government last week issued an advisory to social media platforms underlining the legal provisions that cover such deepfakes and the penalties their creation and circulation may attract.