Pointer Research is a frontier artificial intelligence research organization dedicated to building a clearer, more responsible interface between AI technology and human society.
Artificial intelligence is reconstructing the foundational structures of human society at an unprecedented pace. The capability frontier of large language models is being redrawn on a quarterly basis, autonomous agents are beginning to play decision-making roles in finance, healthcare, law, and scientific research, and the Agentic Web—an internet composed of AI agents—has moved from theory to early prototype.
Yet the challenges brought by this technological revolution far exceed the capacity of existing research institutions to respond. Current AI safety research remains heavily concentrated on the technical dimension of the alignment problem, lacking systematic attention to the societal risks triggered by large-scale AI deployment—concentration of power, the restructuring of labor markets, and the institutionalized propagation of algorithmic discrimination. AI ethics research often devolves into normative manifestos, disconnected from engineering reality and policy mechanisms. The field of economics has yet to establish an analytical framework capable of capturing the new reality of "AI agents as economic actors." The foundational architecture of protocol design and the Agentic Web is almost entirely dominated by a handful of technology companies, lacking an independent voice from neutral academic institutions.
"The true risk of AI is not merely technical loss of control, but institutional aphasia—by the time society is able to name a problem, that problem has already become reality."
Pointer Research was born from this. Our name derives from a simple metaphor: in the future co-authored by humans and AI, we hope to be that pointer—pointing to where the risks lie, pointing to the overlooked margins, pointing to the problems no one has yet named.
Pointer Research is neither an advocate for AI nor an opponent of it. We are clear-eyed observers and active participants in the age of AI. We believe technological acceleration is a fact, not a choice. But we also believe that the direction and consequences of that acceleration depend on humanity's collective choices.
Our research takes human well-being as its ultimate standard, rather than technological progress or economic growth. This means we will sometimes reach conclusions that the tech world or industry would rather not hear.
Pointer Research conducts its work across nine interwoven core directions, each spanning multiple layers of theoretical research, applied research, and community outreach.
Mathematics & Formal Verification
Can AI truly understand mathematical proofs, or is it merely pattern matching? Where are the limits of large language models in formal verification and automated theorem proving? How can AI accelerate mathematical discovery while preserving strict logical reliability? And how does this capability extend to broader domains?
AI Safety & Reliability
How do we define and measure the safety of AI systems in complex, real-world environments? When AI agents hold long-term goals, how do we prevent goal drift and value misalignment? In multi-agent systems, how can local alignment give rise to emergent risks at the system level?
AI Ethics & Value Alignment
Whose values should AI systems be aligned to? How do we handle value conflicts in a pluralistic society? How can AI ethics principles be translated from declarations into operational engineering specifications? In non-Western cultural contexts, what is the path toward localizing AI ethics?
AI Social Impact & Governance
How does large-scale AI deployment reshape the distribution of power? How can regulators implement effective oversight under conditions of technical knowledge asymmetry? What are the pathways through which AI affects democratic institutions, public discourse, and political participation?
Economic Distribution & Labor
How does AI-driven automation affect workers across different skills, industries, and regions? How is the economic value created by AI distributed, and who are the primary beneficiaries? Are existing social safety nets sufficient to withstand the employment shock brought by AI?
Rethinking & Reshaping Education in AI Era
How will AI transform the fundamental ways we learn, teach, and acquire knowledge? Do traditional educational assessment systems remain valid in an era of AI-assisted learning? How can we redesign education systems to cultivate the capabilities truly needed in the age of AI—critical thinking, ethical judgment, and creative thought?
Agentic Web & New Internet
When AI agents become the primary source of internet traffic, how will the architecture and content ecosystem of the web change? How is the agency of human users protected within the Agentic Web? How do we guard against the novel security threats the Agentic Web introduces?
Protocol Design & Interoperability
What protocol standards should govern communication and collaboration among AI agents? How do we design technical protocols that prevent the concentration of power rather than accelerate it? How will the competition between open protocols and closed ecosystems shape the trajectory of AI's development?
Open & Local AI
How can open-source AI models break the monopoly of tech giants and empower the global developer community? What role does localized AI play in privacy protection, cultural adaptation, and knowledge autonomy? How do we build a sustainable open-access ecosystem that ensures the democratization of AI knowledge and tools?