AI

Another principle we adhere to is creating fault-tolerant interfaces. AI-empowered work is not infallible, especially when it comes to generative tasks. Therefore, we design our AI features to allow for easy user corrections. This flexibility ensures that while AI serves as a helpful starting point, the user always has the final say, thus gradually

Des scientifiques de l’Université Carnegie Mellon aux États-Unis, associés à des chercheurs de Microsoft – qu’on ne peut suspecter de technophobie à l'égard de l’intelligence artificielle – viennent de publier une étude qui conclut que plus les personnes utilisent l’IA générative dans leur travail, moins elles feront preuve d’esprit critique. Et qu

In the coming years, as LLM producers consolidate their power, framing alignment as censorship is critical so we can hold producers, and by proxy, the states that influence them, accountable.

Unlike social media platforms, LLM producers argue that they are producing content rather than hosting it. They do this in order to avoid copyright claims. As a result, under many legal jurisdictions, the LLM producer might be held more liable for their models’ outputs than social media platforms would be for the content created by their users.

Information controls introduced now will also snowball into future generations of LLMs. Since new generations of LLMs are reliant on synthetic data generated from previous generations of LLMs, they inherit any model-layer information controls, and it will be harder to reverse any censorship or information manipulation that is baked into earlier mod