Actualités — 2026-09-03

An investigation into Perplexity's search models found that over a third of citations do not contain the numbers they claim, with many pages either inaccessible or lacking the relevant information, raising concerns about the reliability of AI-generated answers.Lire la suite

Research indicates that the widespread adoption of large language models (LLMs) is contributing to a decline in linguistic diversity, as these models often prioritize clarity and concision over creative expression and unique linguistic styles.Lire la suite

ElevenLabs has hired Ashley Kramer, formerly OpenAI's vice president of enterprise sales, as its first chief revenue officer, amid a significant shift in revenue towards enterprise customers. The company's core product is subject to the EU's AI Act, requiring synthetic content marking rules by December 2, 2026.Lire la suite

Despite predictions that self-referentiality would be key to artificial intelligence, recent advancements in AI have achieved conversational intelligence without explicitly building in self-referential abilities. This development has stunned experts like Douglas Hofstadter, who had long believed that self-reference was essential to true intelligence.Lire la suite

Reliance Jio is expanding its cloud PC service, JioPC, to any internet user in India, allowing users to turn aging computers into AI-ready machines without hardware upgrades. The service starts at ₹1,000 for two months and offers up to eight virtual CPUs, 16GB of RAM, and 1TB of storage.Lire la suite

A frustrated job seeker, tired of being ghosted by AI recruiters, decided to turn the tables by using ChatGPT to interview with an AI recruiter, highlighting the absurdity of the modern job hunt and the growing use of AI in the recruitment process.Lire la suite

Google has launched Gemini 3.8 Flash, its latest AI model, which promises to 'work harder' than its predecessor and offer significant improvements in software engineering and autonomous AI agents. The new model is available now for consumers and developers, with introductory pricing starting at $0.75 per million input tokens and $3.75 per million output tokens.Lire la suite

A recent experiment with ChatGPT ads for event seating assignment software yielded disappointing results, with a shockingly low conversion rate of 0.46% and an average time on page of just 7 seconds. The ads' poor targeting is suspected to be the main culprit, with expert analysis suggesting that the traffic generated was of low value.Lire la suite

The rise of generative AI has made much of the internet unusable, with search engines and social media platforms overwhelmed by AI-generated content. As a result, many are rethinking their online habits and seeking out more meaningful interactions.Lire la suite

The increasing reliance on AI agents in software development has led to a concerning trend: experienced engineers are no longer refactoring code, potentially resulting in unmanageable and unmaintainable systems. This shift is attributed to AI agents' ability to navigate complex code without getting lost, thereby removing the human trigger to refactor.Lire la suite

A lawsuit is seeking to force the Trump administration to disclose its framework for testing the safety of AI models, amid concerns that the rules may be biased or inadequate. The lawsuit, filed by Protect Democracy, argues that the public has a right to know how the government is evaluating the risks posed by AI systems.Lire la suite

The Justice Department has filed a statement of interest supporting OpenAI's fair use defense in a copyright lawsuit brought by publishers, including the New York Times, arguing that licensing costs would threaten national security and give foreign adversaries a competitive advantage.Lire la suite

As artificial intelligence transforms industries, US schools are torn between banning AI tools and teaching students to use them responsibly, leaving a growing skills gap that threatens to leave students unprepared for the future.Lire la suite

Meta is ending its tokenmaxxing incentive program, which graded employees on their AI usage, and instead focusing on their overall impact. The company is also encouraging employees to test its new agentic AI tool, Hatch, ahead of a public release.Lire la suite