Des experts en IA incités à renforcer les principes moraux dans leurs recherches

Une nouvelle tribune invite les chercheurs en intelligence artificielle à appliquer activement des lignes directrices éthiques face aux progrès technologiques rapides.

La tribune, publiée par CNET, souligne que les manquements moraux en matière d'IA découlent souvent de principes non activés plutôt que de leur absence. Elle présente six étapes pratiques pour aider les chercheurs à maintenir des normes éthiques, notamment en définissant des lignes rouges personnelles et en résistant aux tactiques de désengagement moral, telles que l'euphémisation du langage et la dilution des responsabilités.

Articles connexes

Tense meeting between US Defense Secretary and Anthropic CEO over AI safety policy relaxation and military access.
Image générée par IA

Pentagon pressures Anthropic to weaken AI safety commitments

Rapporté par l'IA Image générée par IA

US Defense Secretary Pete Hegseth has threatened Anthropic with severe penalties unless the company grants the military unrestricted access to its Claude AI model. The ultimatum came during a meeting with CEO Dario Amodei in Washington on Tuesday, coinciding with Anthropic's announcement to relax its Responsible Scaling Policy. The changes shift from strict safety tripwires to more flexible risk assessments amid competitive pressures.

At the India AI Impact Summit, Prime Minister Narendra Modi described artificial intelligence as a turning point in human history that could reset the direction of civilisation. He expressed concern over the form of AI to be handed to future generations and emphasised making it human-centric and responsible. Experts have warned about risks including data privacy, deepfakes, and autonomous weapons.

Rapporté par l'IA

Members of the Catholic Educational Association of the Philippines said artificial intelligence cannot duplicate the human conscience as they pushed for the responsible integration of AI into the teaching-learning process.

Following IBM's recent findings on AI accelerating vulnerability exploits, a TechRadar report warns that hackers are turning to accessible AI solutions for faster attacks, often trading off quality or cost. Businesses must adapt defenses to these evolving threats.

Rapporté par l'IA

Researchers from the University of Pennsylvania have identified 'cognitive surrender,' where people outsource reasoning to AI without verification. In experiments, participants accepted incorrect AI responses 73.2 percent of the time across 1,372 participants. Factors like time pressure increased reliance on flawed outputs.

Indigenous leaders at the United Nations Permanent Forum on Indigenous Issues are grappling with artificial intelligence's potential to both aid and threaten their traditional lands. While AI tools help monitor deforestation and wildfires, the data centers powering the technology consume vast resources often extracted from Indigenous territories. A new study by former forum chair Hindou Oumarou Ibrahim outlines these opportunities and risks.

Ce site utilise des cookies

Nous utilisons des cookies pour l'analyse afin d'améliorer notre site. Lisez notre politique de confidentialité pour plus d'informations.
Refuser