Latest AI Developments - Topic Citations
Topic ID: 118
Every citation has two independent 0-100 scores. The source Reliability score rates the general reliability of the outlet itself, regardless of the specific claim, based on an eight-tier scale running from primary government and international scientific bodies (CDC, WHO, NASA, IPCC) at the top, through peer-reviewed journals, wire services, mainstream science journalism, and advocacy organizations, down to tertiary references like Wikipedia and unverifiable or absent sources at the bottom. The Accuracy score, by contrast, rates how well a specific citation supports topic statements — a citation from an excellent source can score low if misapplied or can score high if it precisely backs up topic statements.
AI Reinforcing Conspiracy Theories and Climate Denial
- Global WitnessReliability score: 72 | Accuracy score: 85
- Global WitnessReliability score: 72 | Accuracy score: 85
- EuronewsReliability score: 70 | Accuracy score: 74
- EuronewsReliability score: 70 | Accuracy score: 74
Deepfakes Undermine Trust in Authentic Information
- California Law ReviewReliability score: 92 | Accuracy score: 92
- California Law ReviewReliability score: 92 | Accuracy score: 92
Malicious Actors Using AI to Automate Cyberattacks
- AnthropicReliability score: 75 | Accuracy score: 93
- AnthropicReliability score: 75 | Accuracy score: 93
EU Implementing and Adjusting Its AI Framework
- Morgan LewisReliability score: 78 | Accuracy score: 92
- Morgan LewisReliability score: 78 | Accuracy score: 92
AI Invents References
- Scientific ReportsReliability score: 90 | Accuracy score: 85
- Scientific ReportsReliability score: 90 | Accuracy score: 85
- NatureReliability score: 93 | Accuracy score: 92
- NatureReliability score: 93 | Accuracy score: 92
Deepfakes Impersonating Public Figures in Elections
- Federal Communications CommissionReliability score: 95 | Accuracy score: 85
- Federal Communications CommissionReliability score: 95 | Accuracy score: 85
Canadas Consultation on AI Transparency
- Government of Canada (Innovation, Science and Economic Development)Reliability score: 92 | Accuracy score: 95
- Government of Canada (Innovation, Science and Economic Development)Reliability score: 92 | Accuracy score: 95
US Voluntary Commitments and the Enforceable-Oversight Debate
- Al JazeeraReliability score: 78 | Accuracy score: 93
- Al JazeeraReliability score: 78 | Accuracy score: 93
Canadas National AI Strategy
- Government of Canada (Innovation, Science and Economic Development)Reliability score: 90 | Accuracy score: 92
- Government of Canada (Innovation, Science and Economic Development)Reliability score: 90 | Accuracy score: 92
AI Systems Circumventing Intended Restrictions
- TechCrunchReliability score: 74 | Accuracy score: 85
- TechCrunchReliability score: 74 | Accuracy score: 85
AI Models Can Find Software Vulnerabilities
- AnthropicReliability score: 72 | Accuracy score: 92
- AnthropicReliability score: 72 | Accuracy score: 92
AI Detection Remains Imperfect
- Reuters Institute for the Study of JournalismReliability score: 88 | Accuracy score: 90
- Reuters Institute for the Study of JournalismReliability score: 88 | Accuracy score: 90
Competition Driving Gains in Reasoning and Multistep Tasks
- Stanford HAIReliability score: 92 | Accuracy score: 78
- Stanford HAIReliability score: 92 | Accuracy score: 78
Impressive Performance Does Not Guarantee Accuracy
- arXiv / OpenAI (Kalai, Nachum, Vempala & Zhang)Reliability score: 85 | Accuracy score: 90
- arXiv / OpenAI (Kalai, Nachum, Vempala & Zhang)Reliability score: 85 | Accuracy score: 90