Who is Dario Amodei, the Italian-American who scares ChatGPT (but likes Google)- Corriere.it

Dario and Daniela Amodei. Learn these names: they are two Italian-American brothers from San Francisco that all Silicon Valley companies would like with them.

He’s the genius: Stanford, Caltech, Princeton. He studied physics with a keen interest in the functioning of the brain at top American universities. His doctoral thesis at Princeton earned him the Hertz Foundation award. He has worked for Google Brain and Baidu (the Chinese version of Google) where he developed Deep Speech 2, a human speech understanding system awarded as one of the breakthrough technologies in 2016 by MIT in Boston. He has published papers considered milestones in the new artificial intelligence industry, starting with Concrete problems in AI Safety. Her sister Daniela (the father’s name is Riccardo), follows close behind. She has always worked with him. You also often co-signed the same papers.

(Here the episode of the podcast Geni Invisibili by Massimo Sideri dedicated to Artificial Intelligence with Maria Chiara Carrozza of the Cnr and Giorgio Metta of the Iit, here the complete series)

The doubts that remain

But what is so special about Dario Amodei? In line with Silicon Valley mantras, he doesn’t care about wealth (I don’t think it’s a good motivation and indeed it tends to change your plans and goals). A non-secondary detail, as we shall see. Because Dario is concerned about the entry of artificial intelligence into our society. His role in ChatGPT was Ai safety research, AI reactions manager. And he hadn’t liked the massive investment of big tech as early as 2020 (last week Microsoft added 10 billion in OpenAi).
In essence Amodei is developing Isaac Asimov’s new laws of robotics but for artificial intelligence. Though virtually unknown to the general public, until now, he is a well-respected name in the Bay Area community of Ai experts. But with an original approach. In one of the interviews released in recent years, Amodei recounted an episode that occurred when he was in Google Brain: a Google face recognition system had studied many images, with the usual bias: too many Caucasians. Among the images the developers had also introduced those of some primates. Result: starting from the color, the system had begun to categorize the photographs of black people as “gorillas”.
“Artificial intelligence is not racist” explained Amodei. It doesn’t have enough data. In short, more a problem of “ignorance”, in the original sense of the term, to be ignored. Whether it is artificial or human intelligence, a probably apocryphal aphorism attributed to Albert Einstein is obviously always valid: “The mind is like a parachute. Only works if opened”. Amodei is an optimist and makes no secret of it (“human intelligence can win the challenge”). Google evidently too. We just have to put ourselves on the waiting list to also test Claude, the anti-ChatGPT.