Indonesia AI Community Facebook group. 6K members.

Mabel AI's inclusive interpretation technology highlighted in MIT Technology Review! 🚀 🚀 🚀 

We are extremely proud to feature -- as the only company -- in a comprehensive overview on efforts to improve the diversity of AI voices. Moreover, we are honored to support the Mozilla Common Voice project, as we share the same mission of inclusive speech AI. 🌍 🌎 🌏 

At Mabel, we develop our own AI models without relying on any third-party APIs -- a prerequisite in healthcare and other public services, where confidentiality is essential. This also means that we can deploy our solution fully on-premise and even on-device. 🔒 

For that reason, representative, ethically sourced voice data is a critical resource for us, as well as for other AI-native, social impact companies. We are thrilled to see this important issue, and our own efforts, highlighted in one of the world’s leading technology magazines! 🌟 ✨ 💫

Generative AI, particularly ChatGPT, creates new managerial concerns. This article addresses a crucial managerial challenge: While employees are including the tool as their private knowledge assistant in many aspects of their daily work, it is difficult for managers to see and understand the impact within their organization. There is more and more literature on what ChatGPT means for business, but we cannot say much as long as we are uninformed about its actual use. So, getting an insider’s perspective is needed to think about the consequences for organizations. The article is grounded in a qualitative study of employees’ experiences of interacting with ChatGPT among 50 early ChatGPT adopters. First, we show how employee-ChatGPT relations develop from private, experimental use into an integral part of knowledge work. Second, we identify six affordances: searching for information, brainstorming ideas, structuring content, writing a first draft, embellishing text, and proofing work. Further, we highlight three looming side effects of ChatGPT use that threaten knowledge ties, the quality of knowledge in organizations, and how work roles are configured. Accordingly, this article offers managers guidance in managing a ChatGPT-empowered workforce in a way that aims to mitigate these side effects while harnessing opportunities.

En federal domstol i USA har avvisat en stämningsansökan där nyhetssajterna Raw Story Media och AlterNet stämt OpenAI. Företagen hävdade att OpenAI använt deras artiklar utan tillstånd för att träna ChatGPT.

Domaren avslog fallet eftersom sajterna inte kunde visa någon konkret skada. Hon menade också att ChatGPT skapar nytt innehåll baserat på inlärda mönster, snarare än kopierar text direkt.

Jag skulle säga att även om amerikanska domar inte har någon juridisk verkan i Sverige så är domarens resonemang intressant. Särskilt argumentet att AI genererar nytt innehåll istället för att kopiera befintligt material.

Det ska bli intressant att följa utvecklingen här i Sverige eftersom vi står inför liknande frågor om AI och upphovsrätt i Sverige och EU.

https://www.bloomberglaw.com/public/desktop/document/RawStoryMediaIncetalvOpenAIIncetalDocketNo124cv01514SDNYFeb282024/4?doc_id=X18AGVQ2BEU9DCBFDQULD4SNTVV