The head of the ChatGPT developer OpenAI believes there will in future be more content created by artificial intelligence than by humans.
Before this happens, society needs to prepare for the negative consequences of the AI boom, Sam Altman said at an event organised by chip giant Intel.
Altman foresees potentially negative effects on elections in the foreseeable future. His comments come as AI-generated footage, or deepfakes, are already being weaponized ahead of India’s elections later this year.
The risks associated with cybersecurity and bioweapons must also be taken more seriously as AI develops, Altman argued.
“This is not going to be only a good story, but it’s going to be a net good story,” Altman said.
With the help of AI, it will be possible to accelerate scientific research, cure diseases and improve education, Altman emphasised, saying it is difficult to imagine today how much better the future will be.
And yet a regulatory role for governments is needed even more than with other technologies, particularly now while the AI models are still relatively weak.
Altman said a company like OpenAI should not be able to secretly develop an AI that is smarter than humans and then suddenly unleash it on the world. Society and its institutions must be given the time to gradually adapt to the development.
OpenAI’s ChatGPT, which kicked off the hype surrounding AI just over a year ago, is trained, like its rival chatbots, on huge amounts of information and can formulate texts at the linguistic level of a human, write software code and summarize information.
Voices in various industries have been forecasting major shifts in everyday practices during the pending “AI era”.
However one disadvantage of AI-based text generators, which estimate word for word how a sentence should continue, is that the software sometimes gives completely wrong answers, even if it was based on correct information.