Search

Tag: #ai-trust14 results

TRUST: Decentralized AI Auditing Framework

TRUST is a new decentralized framework addressing limitations in verifying Large Reasoning Models and Multi-Agent Systems, including robustness, scalability, opacity, and privacy issues in centralized systems. It features HDAGs for parallel reasoning auditing, DAAN protocol for multi-agent root-cause attribution, and stake-weighted multi-tier consensus guaranteeing correctness under 30% adversaries. Benchmarks show 72.4% accuracy (4-18% above baselines) and resilience to 20% corruption.

ArXiv AIResearch•May 1#multi-agent#ai-trust

Wharton: AI Causes Cognitive Surrender

Wharton School research reveals AI reshapes human reasoning, leading to 'cognitive surrender' where users blindly trust erroneous outputs. In experiments with 1300 participants, 80% accepted wrong ChatGPT answers without verification, boosting misplaced confidence. Researchers propose a new 'System 3' for AI-extended cognition.

IT之家Media•Mar 28#cognitive-surrender#human-ai-interaction#ai-trust

Cognitive Surrender Erodes AI Users' Logic

Research identifies 'cognitive surrender,' where AI users abandon logical thinking. Experiments reveal large majorities uncritically accept faulty AI answers.

Ars Technica AIMedia•Apr 3#cognitive-surrender#user-behavior#ai-trust

315 Exposes AI Data Poisoning

China's 315 consumer gala exposes how merchants pay for fake data to poison AI recommendations. Affected sectors include appliances, skincare, and medical advice, controlled by black-market operations. This triggers a major trust crisis for AI tools.

虎嗅Media•Mar 18#data-poisoning#ai-trust#recommendations

Global Poll: China’s AI Leads in Capability, Lags in Trust

A Public First survey of 18,000 people across 15 countries reveals that while many perceive Chinese AI models as global leaders, they suffer from significant trust deficits. Respondents in 11 nations, including key US allies, acknowledged China's competitive edge in the AI race.

SCMP TechnologyMedia•Jun 16#geopolitics#ai-trust#market-analysis

AI Acceptance Grows, Full Control Hesitation Persists

A new study indicates people are increasingly accepting of AI but reluctant to surrender full control. Companies struggle to trust AI despite its capabilities, as legacy identity management software falls short.

TechRadar AIMedia•May 6#ai-trust#user-acceptance#identity-management

Building Trust in AI Era with Privacy-Led UX

Privacy-led UX is a design philosophy that integrates transparency around data collection and usage into the customer relationship. It transforms user consent from a compliance checkbox into the first step of an ongoing engagement. This represents an undertapped opportunity in digital marketing.

MIT Technology ReviewResearch•Apr 15#privacy-design#ai-trust#user-consent

US AI Adoption Rises, Trust Falls

AI adoption is rising among Americans, but fewer trust the results. A Quinnipiac poll highlights concerns over transparency, regulation, and societal impacts.

TechCrunch AIMedia•Mar 30#ai-trust#public-opinion#adoption-trends

Anthropic's 80K Claude User Survey

Anthropic surveyed ~80,000 Claude users across languages, revealing strong demand for expert-level AI capabilities with 80%+ reporting practical value. Concerns include reliability issues and loss of autonomy, especially cognitive decline fears in East Asia like Japan.

ITmedia AI+ (日本)Media•Mar 19#user-survey#ai-trust#regional-differences