AI Expert Calls for a Shift to Ethical AI Research

Timnit Gebru, the founder of the Distributed Artificial Intelligence Research Institute, called for scholars to employ more ethical approaches in artificial intelligence research at an event hosted Tuesday by the Radcliffe Institute for Advanced Study.

During her talk, part of a Radcliffe lecture series on artificial intelligence, Gebru shared her vision for interdisciplinary AI and her calls for changes in academia. Gebru delivered a virtual talk to attendees, followed by a Q&A with Himabindu “Hima” Lakkaraju, an assistant professor at Harvard.

The conversation began with an introduction to DAIR, which works with researchers from different backgrounds to conduct research around the world. Gebru said traditional research practices can often be “exploitative.”

“We see this a lot, where we have people in one community that have specific types of knowledge, and then people in the academic and research community that go to this community, and without supporting them or collaborating with them, [take] this knowledge,” Gebru said.

“And this group of people whose knowledge was extracted from are not getting any of [the] benefits,” she added.

She spoke about “the publishing rat race” — expectations in academia that researchers must publish lots of work in short periods of time. Gebru said her institute seeks to create a more sustainable working environment for researchers.

“We have a limit, and we want to be healthy and thriving researchers, and we can’t do our best work if we’re not healthy and thriving,” she said.

Before founding DAIR, Gebru was a co-lead of the Ethical AI working group at Google. In December 2020, Google fired Gebru over controversy concerning a paper the working group published.

During the Q&A portion of the event, Gebru focused on her vision for the future of DAIR and AI research. She said she hopes to build “a thriving group of researchers who are healthy, who are embedded in their communities.”

Read more

Mots-clés : cybersécurité, sécurité informatique, protection des données, menaces cybernétiques, veille cyber, analyse de vulnérabilités, sécurité des réseaux, cyberattaques, conformité RGPD, NIS2, DORA, PCIDSS, DEVSECOPS, eSANTE, intelligence artificielle, IA en cybersécurité, apprentissage automatique, deep learning, algorithmes de sécurité, détection des anomalies, systèmes intelligents, automatisation de la sécurité, IA pour la prévention des cyberattaques.