#TechnologyPolicy

Recent research from the University of California San Diego and University of Wisconsin–Madison reveals critical insights about public expectations for algorithmic decision-making in high-stakes contexts. The study, presented at the 2025 ACM CHI conference, explored how ordinary people react when multiple high-accuracy machine learning models reach different conclusions for identical applications. The findings challenge both current industry practices and academic assumptions about fair automated decision-making, with direct implications for Thailand’s rapidly expanding use of AI systems in financial services, employment, and government programs.

A new study from researchers at the University of California San Diego and the University of Wisconsin–Madison, presented at the 2025 ACM CHI conference, examines how the public wants decisions made when multiple high-accuracy AI models disagree. The findings are especially relevant to Thailand as AI use grows in finance, employment, and government services.

The study centers on multiplicity—the reality that many models can achieve similar accuracy but still produce different predictions for the same case. This raises ethical questions for organizations choosing which model to deploy, particularly for loans, jobs, or social services. Data resonates with Thailand’s push to sharpen AI risk management guidelines in finance, signaling regulators’ attention to fairness in automated decisions.

A new study by computer scientists at the University of California San Diego and the University of Wisconsin–Madison warns that relying on a single “best” machine learning (ML) model for high‑stakes decisions — from loan approvals to hiring — can undermine perceived fairness, and that ordinary people prefer human arbitration when equally good models disagree. The research, presented at the 2025 ACM CHI conference, explored how lay stakeholders react when multiple high‑accuracy models reach different conclusions for the same applicant and found strong resistance to both single‑model arbitrariness and to solutions that simply randomize outcomes; instead participants favored wider model searches, transparency and human decision‑making to resolve disagreements UC San Diego report and the authors’ paper Perceptions of the Fairness Impacts of Multiplicity in Machine Learning (CHI 2025) presents the detailed results and recommendations.

A new era of digital inequality is emerging worldwide, as recent research reveals that artificial intelligence (AI) computing power is rapidly becoming concentrated in just a handful of countries and companies. This growing divide between AI “haves” and “have-nots” is set to disrupt economies, fuel geopolitical rivalries, and shape the scientific and social futures of entire nations, according to a major new investigation by Oxford University researchers and reporting by The New York Times (nytimes.com).

A new wave of digital inequality is forming as AI computing power concentrates in a few countries and firms. An Oxford University study, reinforced by in-depth reporting from a leading U.S. newspaper, shows that most powerful AI systems run on data centers owned by a handful of players. This gap threatens economic competitiveness, scientific progress, and national security for countries outside the core hubs, including Thailand.

The opening of OpenAI’s planned massive data center in Texas illustrates the scale of resources now required to run cutting-edge AI. In contrast, researchers in some regions operate aging hardware in makeshift facilities, underscoring a widening gulf in compute power that is outpacing growth elsewhere.

A rapidly escalating wave of academic dishonesty has gripped universities worldwide, with a recent UK study exposing a dramatic rise in students caught cheating with artificial intelligence tools like ChatGPT. The findings, which reveal nearly 7,000 proven cases of AI-facilitated cheating between 2023 and 2024, spotlight an urgent challenge for educators not just in the UK but across the globe, including Thailand. Experts warn these figures are likely just the “tip of the iceberg,” suggesting that the true scope of technology-driven misconduct is far greater and largely undetected—potentially transforming how societies view and manage academic integrity (The Guardian).

A surge in AI-assisted cheating is rattling universities worldwide, with a UK study showing a sharp rise in cases involving tools like ChatGPT. The research reports nearly 7,000 proven instances of AI-facilitated cheating in 2023–2024, signaling a widening challenge for educators beyond the UK and into Thailand. Experts warn the figures may only scratch the surface, as advanced AI capabilities outpace detection methods.

The trend marks a shift from traditional cheating to high-tech shortcuts that deliver rapid and sophisticated results. The latest findings indicate an average of 5.1 students per 1,000 engaged in AI-enabled cheating in 2023–24, up 219 percent from the previous year. Projections suggest this rate could rise further, underscoring the expanding scale of the problem.

A groundbreaking study by Harvard University has found that GPT-4o, OpenAI’s latest large language model, exhibits behaviors akin to human cognitive dissonance—a psychological phenomenon previously thought to be exclusively human. The findings, reported in the prestigious Proceedings of the National Academy of Sciences on May 28, 2025, raise fresh questions about how advanced AI systems process information and make decisions, carrying significant implications for Thailand’s growing embrace of AI-driven technology in education, health, and society at large (TechXplore).

A recent Harvard-led study suggests that GPT-4o, OpenAI’s newest large language model, exhibits behaviors resembling human cognitive dissonance. Published in the Proceedings of the National Academy of Sciences on May 28, 2025, the findings prompt fresh questions about how advanced AI processes information and makes choices. For Thailand, where AI is increasingly used in classrooms, clinics, and public services, the study raises important considerations for safeguarding reliability and trust in AI-powered tools.

In an event reverberating across the scientific community, the Massachusetts Institute of Technology (MIT) has publicly rescinded its institutional support for a recent artificial intelligence (AI) research paper. The move, reported by Retraction Watch, raises fundamental questions about the future of AI-authored academic articles and the ethical challenges facing research institutions worldwide.

The decision by MIT, a global powerhouse in science and technology innovation, has significant implications for the credibility and accountability of scientific literature. AI-generated content, increasingly prevalent in both drafting and data analysis duties, is under the microscope for issues relating to originality, transparency, and academic integrity. For Thai academic institutions, researchers, and the broader public, this case exemplifies the urgent need to establish clear policies and ethical guardrails as the adoption of AI tools accelerates throughout the research landscape.

A move with wide reverberations in science circles: MIT has publicly withdrawn its institutional support for a recent AI-driven research paper. The decision, reported by Retraction Watch, highlights ongoing questions about AI’s role in drafting and analyzing scientific work and the ethical responsibilities of research institutions worldwide. For Thai readers, the episode signals a crucial moment to establish clear policies on AI use in research and publication.

MIT’s reversal raises questions about how much credit AI should receive in scholarly work and whether automated systems can or should participate as contributors or co-authors. The cited paper reportedly used advanced language models to generate substantial portions of text and analysis, prompting concerns that automation could bypass essential peer review steps or recycle unverifiable content. A technology ethics expert noted that while AI can summarize or rephrase existing research, it cannot reliably assess the validity of scientific claims or ensure precise citations.

A groundbreaking wave of neuroscience research is redefining what it means to think—and, crucially, why artificial intelligence (AI) still falls far short of the intricacies of the human mind. A newly published feature in Salon highlights the distinct evolutionary adaptations that make the human brain more than a glorified computer, challenging decades-old assumptions fundamental to AI development and the neural network concept that underlies machine learning models.

For years, popular understanding—and much of AI research—has treated the brain as a vast network made up of nearly identical neurons whose intelligence emerges through the patterns of their collective firing. This view inspired so-called artificial neural networks, computer systems designed to solve problems by mimicking the way brains process information. Such analogies, cemented over decades, fostered the belief that if machines could imitate the structure and connectivity of brains, they might one day match, or even surpass, human intellect. But recent scientific discoveries show this metaphor misses the mark in fundamental ways, with profound implications for both neuroscience and the future of AI.

New neuroscience findings are reshaping what we mean by “thinking.” They show that artificial intelligence, though powerful, still lags far behind the human brain’s complexity and adaptability. A recent feature highlights how evolutionary advances give humans unique capabilities that machines struggle to replicate, challenging long-standing AI assumptions rooted in neural network models.

Why this matters for Thai readers. As Thailand accelerates digital transformation in health, education, and business, understanding how intelligence works—biological and artificial—helps shape better policies and practical AI applications. These insights also matter for how AI is used in Thai classrooms, hospitals, and public services, where accuracy, empathy, and cultural context matter.