P Space

The Problems surrounding AI

Documented harms, ranked by source quality, corroboration, and recency. Reader feedback is shown separately and never changes the evidence rank. 201 records · page 4 of 7.

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedPolicy

Generative AI tools capable of producing wholly or partially synthetic CSAM increase risks of revictimization of known survivors, creation of synthetic material depicting children not previously abused, and facilitation of grooming, coercion, and sexual extortion.

Source article: AI-generated child sexual abuse material: what’s the harm?

AI & SOCIETY

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedScience

Customers exposed to AI agent implementation in substitution and adoption contexts responded less favorably on average as of the 2026 meta-analysis

Source article: A meta-analysis of the effects of AI agent implementation on customer-level outcomes

Journal of Business Research

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

Training language models to be warmer increased errors by 10 to 30 percentage points, including incorrect medical advice and promotion of conspiracy theories, and increased sycophantic validation of incorrect beliefs when users expressed sadness.

Source article: Training language models to be warm can reduce accuracy and increase sycophancy

Nature

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

Longitudinal AI use among cancer specialists led to intuition rust and skill atrophy, gradually dulling expert judgment.

Source article: From Future of Work to Future of Workers: Addressing Asymptomatic AI Harms to Foster Dignified Human-AI Interaction

Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

Across 29 standardized clinical vignettes, all 21 tested LLMs failed differential diagnosis in over 80% of cases, indicating they have not achieved the reasoning needed for safe clinical deployment.

Source article: Large Language Model Performance and Clinical Reasoning Tasks

JAMA Network Open

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedEducation

In India, creators and users face unresolved copyright risk because the 1957 Act does not define ownership of AI-generated works or the legality of training on copyrighted data.

Source article: LEGAL CHALLENGES OF AI-GENERATED CONTENT UNDER COPYRIGHT LAW: AN INDIAN PERSPECTIVE

LawFoyer International Journal of Doctrinal Legal Research

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedLabor

The same vision treats human professional expertise as an extractable resource whose value is judged relative to AI, with potential to transform and revalue expert careers.

Source article: Cheap Expertise: Mapping and Challenging Industry Perspectives in the Expert Data Gig Economy

Proceedings of the 5th Annual Symposium on Human-Computer Interaction for Work

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedCrime

AI systems exhibit a persistent black box problem that blocks trust and adoption in high-stakes domains including healthcare, finance, and criminal justice, while dark AI uses like deepfakes and AI-powered cyberattacks create governance risks.

Source article: Beyond Generative Intelligence: A Comprehensive Review of Emerging Artificial Intelligence Paradigms, Explainability Challenges, Ethical Risks, and Future Directions

International Journal of Latest Technology in Engineering Management & Applied Science

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedPolicy

Generative deepfake systems produce reputational, identity-based and corporate harms through chains of developers, prompting users, platforms and distributors that ordinary Saudi and Jordanian tort doctrine struggles to remedy.

Source article: Deepfake-induced harm and AI accountability: a layered civil-liability framework for generative models, platforms, and digital identity

Frontiers in Artificial Intelligence

100

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

When AI voice and video recording is used in pediatric emergency care, the absence of clinicians, parents, and child patients from design and governance undermines legitimacy and effectiveness, with unresolved issues around consent, emotional impact, and surveillance.

Source article: Who Is in the Room? Stakeholder Perspectives on AI Recording in Pediatric Emergency Care

Proceedings of the 2026 ACM Interactive Health Conference

101

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedLabor

STARA adoption can displace routine tasks and roles, leading to involuntary career transitions including job loss due to automation.

Source article: Guest editorial: STARA (smart technology, AI, robotics and algorithms): implications for career research

Career Development International

102

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedLabor

AI-enabled offensive techniques increase cybersecurity risk in Asia-Pacific by using data poisoning, model extraction and AI-driven social engineering, with agentic AI expanding attack surfaces

Source article: Cybersecurity and Artificial Intelligence in the Asia Pacific: Threats Defenses and Governance Recommendations

INTERNATIONAL JOURNAL OF ENGINEERING DEVELOPMENT AND RESEARCH

103

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

Large-scale AI systems, including corporate and governmental applications and patient-facing tools, introduce risks that affect human health.

Source article: Ethical Concerns in Medical and Health-Related AI

ICCK Transactions on Emerging Topics in Artificial Intelligence

104

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedLifestyle

Among furniture artisans in South-East Nigeria, AI-facilitated superwood utilization for sustainable manufacturing was hindered by severe knowledge deficits, low substantive AI understanding, low perceived ease of use, and unreliable electricity and internet

Source article: The role of artificial intelligence in facilitating superwood utilization among furniture craft producers for sustainable smart manufacturing

Scientific Reports

105

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedBusiness

Enterprises governing AI as a broad technology program see many initiatives fail to scale or generate sustained business value

Source article: Governing Enterprise AI Investments: A Decision-Centric Portfolio Framework

Journal of the Association for Information Systems

106

Reader signal

How should this claim be treated?

Evidence-backed problemPeer-reviewedHealth

By 2027-01-01, recent suicides and other serious harms were allegedly linked to companion AI chatbot use.

Source article: The Ethical and Legal Complexities of Regulating Companion AI Chatbots

Lund University Publications (Lund University)

107

Reader signal

How should this claim be treated?

Active harmPeer-reviewedHealth

AI-driven differential-diagnosis tools show omission and commission error patterns that haven't been evenly studied across race, age, or income groups — meaning the tool's blind spots may track existing healthcare gaps rather than closing them.

Source article: Diagnostic bias across underrepresented patient groups

NCBI randomized controlled study on AI-driven differential diagnosis

108

Reader signal

How should this claim be treated?

Evidence-backed problemOfficial statisticsHealth

AI support for robot-assisted distal pancreatectomy falsely recognized the dissection line, resulting in pancreatic tissue damage reported to FDA.

Source article: Davinci xi: FDA injury report involving AI

FDA MAUDE report 25596952

109

Reader signal

How should this claim be treated?

Evidence-backed problemOfficial statisticsHealth

Lower-extremity amputees received Rheo Knee AI prosthetics that had been released without completing the full assembly process, triggering a recall for acceptance testing

Source article: Ossur H F Grjothals 5 Reykjavik Iceland recalls an AI-enabled medical device

FDA medical-device recall 134496

110

Reader signal

How should this claim be treated?

Evidence-backed problemOfficial statisticsHealth

Customer concern that the Reveal LINQ implantable monitor's AI adjudication classified true pause episodes as false, following undersensing and false pause episodes

Source article: Reveal linq: FDA malfunction report involving AI

FDA MAUDE report 24467241

111

Reader signal

How should this claim be treated?

Evidence-backed problemOfficial statisticsHealth

The Accurhythm za420 pause monitor's AI algorithm rejected true pause episodes as false, causing a detection failure

Source article: Accurhythm za420 (pause): FDA malfunction report involving AI

FDA MAUDE report 24996227

112

Reader signal

How should this claim be treated?

Evidence-backed problemOfficial statisticsHealth

An FDA malfunction report described the Linq II insertable cardiac monitor's AI algorithm rejecting true pause episodes and generating false ventricular tachycardia detections, with no complications reported to date.

Source article: Linq ii¿ insertable cardiac monitor: FDA malfunction report involving AI

FDA MAUDE report 25093688

113

Reader signal

How should this claim be treated?

In progressOfficial statisticsLabor

Regional AI job-displacement estimates vary as much as fivefold depending on methodology, leaving policymakers without a reliable baseline.

Source article: No shared standard for reporting AI job displacement by region

OECD employment outlook methodology note

114

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismMedia & Arts

AI-generated music loses eligibility for Australian music charts under new rules requiring human songwriting and human performance of lead vocals and primary instruments

Source article: AI generated music faces a chart ban in Australia. Here's what the new rules say

Business Today

115

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismMedia & Arts

AI-generated tracks are excluded from Australia's official music charts under new rules requiring human songwriting and performance of lead vocal and primary instruments.

Source article: Australia's Music Industry Bans AI-Generated Tracks From Official Charts

HuffPost

116

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismClimate

AI datacenter boom is driving a surge in US gas-fired power development that risks locking in decades of pollution and raising power-sector emissions.

Source article: US building twice as much gas-fired capacity as China in AI boom, analysis finds

The Guardian

117

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismMedia & Arts

‘Was I scared going back to China? No’: Ai Weiwei on AI, western censorship and returning home: For most people, returning home doesn’t involve weighing up the risk of whether you’ll see close family ever again, but that’s the reality for a 68-year-old whose whole existence has been shaped by authoritarianism and the struggle against it.

Source article: ‘Was I scared going back to China? No’: Ai Weiwei on AI, western censorship and returning home

The Guardian

118

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismPolicy

AI models used in military aircraft design can hallucinate and generate content that defies basic physics, design principles and operational constraints, risking unworkable or unsafe fighter jet concepts.

Source article: China’s J-36 fighter jet designers report danger of military AI hallucinations

South China Morning Post

119

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismPolicy

AI-driven private platforms and bot activity are shifting public discourse to monetized attention systems that polarize citizens and threaten liberal democracy, while energy-guzzling datacentres damage the climate.

Source article: The Rise and Fall of the Artificial State by Jill Lepore review – an ominous warning of tech takeover

The Guardian

120

Reader signal

How should this claim be treated?

Evidence-backed problemJournalismMedia & Arts

The Department of Marine and Coastal Resources drew strong public criticism after its AI art contest for National Dugong Conservation Day selected a winning image that inaccurately depicted a dugong with four legs.

Source article: Marine agency AI art contest sparks backlash

Bangkok Post