NASA’s new director of artificial intelligence, David Salvagniniwarned this Wednesday that the results offered by this technology must be approached with a critical attitude, because there may be “errors of omission” that affect its responses.
“There are times that it generates a false response that may be obvious and easier to detect. I would be more concerned about errors of omission,” he says.
OpenAI abolishes its AI risk analysis department
“What happens when you give an answer but there was a large amount of data that you didn’t really refer to?” asks Salvagnini, who urges a detailed analysis of the elements that have led to those answers.
“Do we understand the origin of that data? Do we trust its accuracy? Because if the answer to those questions is negative, our confidence in the AI result should be discarded or reduced,” he said.
“I would like artificial intelligence not to be called that, but to be known as digital help”
Salvagnini He was appointed on May 13 as NASA’s first AI director to align the strategic vision and planning of the use of that technology in the US space agency.
Almost half of companies plan to reduce their workforce in a year due to AI
The new position responded to the Joe Biden Administration’s October 2023 executive order on the development and safe use of AI, which called on all federal agencies to create such a position.
“AI is not responsible for the result, people are. We have to be aware of our responsibilities. I would almost like artificial intelligence not to be called that, but to be known as digital help, like a digital assistant that one has access to receive help in the decision-making process,” he concluded.
