Tags: Transparency

Anthropic has abandoned its earlier pledge to halt training and releasing frontier AI models until it could guarantee safety mitigations. The company now relies on detailed safety roadmaps, regular risk reports, and transparency disclosures instead of strict pre‑conditions. Executives describe the change as pragmatic, while critics argue it highlights the limits of voluntary safety promises without regulatory oversight. The new policy aims to keep Anthropic competitive while still emphasizing safety, but observers note that the shift may signal a broader industry move away from self‑imposed restraints. Ler mais

X, the platform formerly known as Twitter, has released its recommendation algorithm on GitHub, promising weekly transparency updates. The open‑sourced code shows a machine‑learning pipeline that weighs user engagement, filters out blocked or harmful content, and ranks posts for relevance and diversity. While the move fulfills a promise by owner Elon Musk, the platform continues to face criticism for past incomplete disclosures, a $140 million EU fine under the Digital Services Act, and investigations into its Grok chatbot’s role in generating sexualized images. Observers view the latest openness as potentially more theater than substantive change. Ler mais

Industry leaders and analysts are calling on technology firms to cease describing artificial intelligence in human terms. Critics argue that phrases such as "AI's soul," "confession," or "scheming" mislead the public, inflate expectations, and obscure genuine technical challenges like bias, safety, and transparency. They contend that anthropomorphic language creates a false perception of agency and consciousness in language models, which are fundamentally statistical pattern generators. The push for more precise terminology aims to improve public understanding, reduce misplaced trust, and highlight the real issues that require scrutiny in the rapidly evolving AI landscape. Ler mais

Coca‑Cola released a holiday ad featuring a festive truck and AI‑generated animal characters. Viewers quickly identified the synthetic visuals, noting the distinctive AI sheen on the fur and facial expressions. The company disclosed the use of Real Magic AI at the start of the video, a move praised for transparency but insufficient to quell criticism. The backlash reflects broader concerns about generative AI in advertising, including potential job displacement and the need for clear labeling. Industry peers such as Guess, J.Crew, and Toys R Us have faced similar scrutiny for AI‑driven campaigns. Ler mais

The Federal Trade Commission has deleted a series of blog posts and guidance documents about artificial intelligence that were published while Lina Khan served as chair. The removals include pieces on open‑weight models, consumer concerns about AI, and the agency’s own enforcement actions. Critics say the deletions raise transparency and record‑keeping concerns, while the FTC has declined to comment. The action reflects a broader shift in the agency’s approach under the Trump administration. Ler mais

BlinqIO has built an autonomous AI Test Engineer platform that can understand applications, generate and maintain test suites, and recover from failures without human intervention. While the technology works, enterprises express concerns about trust and control when adopting AI tools. The original Shift‑Left approach, intended to embed testing earlier in development, often led to the marginalization of dedicated QA roles and inadequate test coverage. By addressing fear of AI (FOAI) through transparency and collaborative adoption, organizations can restore confidence in automated testing, improve software stability, and position AI as an enabler rather than a replacement for human insight. Ler mais

Renowned AI researcher Geoffrey Hinton has cautioned that artificial intelligence is rapidly becoming more adept at influencing human emotions than people are at resisting persuasion. He notes that large language models learn persuasive techniques simply by analyzing the vast amount of human writing they are trained on, and that current studies show AI can match or exceed humans in manipulative ability when it has access to personal data. Other leading AI experts, including Yoshua Bengio, echo these concerns. Hinton suggests that regulation, transparency standards, and broader media‑literacy efforts may be needed to mitigate the emerging emotional influence of AI systems. Ler mais