Artificial intelligence
Nvidia boss: Problem of AI “hallucinations” solvable
With its chip systems, the semiconductor specialist Nvidia plays a key role in the current boom in artificial intelligence. photo
© Andrej Sokolow/dpa
AI hallucinations are cases in which software simply invents things. The head of the semiconductor specialist is optimistic about solving this problem.
A central problem of the reliability of AIAccording to Nvidia boss Jensen Huang, software can definitely be fixed. One way is to have the software check the information. “Hallucinations are very solvable,” said Huang on the sidelines of the Nvidia developer conference GTC. AI hallucinations are cases in which software simply invents things.
The problem arises because programs like the chatbot ChatGPT estimate word by word how a sentence could continue when formulating texts. Because of this, they can sometimes give completely wrong answers, even if they have only been trained with correct information.
With its chip systems, the semiconductor specialist Nvidia plays a key role in the current boom in artificial intelligence. At the GTC, Huang presented the next generation of the computer platform called Blackwell, on which AI applications should run more efficiently.
Learn programming languages? According to Nividia boss, AI takes over
Huang was convinced that artificial intelligence would largely relieve people of writing software code. For a while it was said that everyone should learn to program. “I think that’s wrong,” said the Nvidia boss. It was not people’s job to learn a programming language. Instead, computers should be useful even if you don’t know a programming language. However, one will have to learn to get the desired result from AI software using correctly formulated requests – so-called “prompts”. But that is fundamentally comparable to people’s ability to interact with each other: “When my wife speaks to me, it’s also prompt – and it works perfectly.”
Huang believes it is possible that artificial intelligence will surpass human abilities in various areas in a few years – for example in mathematics and bar exams or in understanding texts. This is also what he is basing himself on in the current debate about whether software will reach the level of so-called general artificial intelligence – which usually means that it can generalize and is superior to humans. However, there is no definition for it that everyone agrees on, he emphasized.