Cognisee Convenes 'Superintelligence For Humanity' Summit To Advance Artificial Collective Intelligence

(MENAFN- EIN Presswire) EINPresswire/ -- Cognisee, a frontier AI lab building Artificial Collective Intelligence, today announced the conclusion of Superintelligence for Humanity, the first invitational summit in a new series focused on AI dignity, sovereignty, and accountable human-AI systems.

Held May 18–19, 2026, at Harvard and the California Institute for Machine Consciousness, the summit brought together frontier researchers, AI builders, governance experts, and institutional leaders to examine a question increasingly central to the future of AI: as systems move from generating content to shaping decisions, institutions, and physical environments, what infrastructure is needed to preserve human agency, provenance, accountability, and meaning?

The summit advanced Artificial Collective Intelligence, or ACI, as Cognisee's research direction for AI systems that learn and coordinate across humans, agents, institutions, and knowledge communities. Rather than treating intelligence as the output of a single centralized model, ACI explores how intelligence can emerge across governed networks while preserving tacit expertise, institutional memory, consent, authorship, and human agency.

“The history of intelligence on Earth is not only a history of bigger and bigger brains. It is a history of new forms of coordination. Intelligence is rarely monolithic - it is distributed, adaptive, embodied, and constantly evolving,” said Dr. Olaf Witkowski, co-founder of Cognisee PBC and founding director of Cross Labs, who convened the summit.“ACI is not a finished answer. It is a research agenda for understanding how intelligence scales without losing diversity, agency, memory, and meaning.”

The summit was designed less as a conventional conference than as a working salon. Across two days, participants from fields that rarely meet in the same room - AI architecture, artificial life, collective intelligence, security, consciousness, linguistics, cultural memory, and institutional governance - explored architectures beyond scale, adaptive learning, multi-scale minds, and sovereign knowledge preservation.

Rather than settling questions, the summit sharpened them. Participants debated whether AI most needs new capabilities or better use of what already exists - contrasting those who see untapped potential waiting to be connected to scientists and communities with those who argue that continual learning remains an unsolved and necessary frontier. A second tension focused on how much discovery AI should drive on its own, and how much human agency should remain as systems begin to explore problem spaces faster than people can follow. A final question asked whether governance, benchmarks, provenance, and accountability can be added to open-ended intelligence without diminishing the curiosity and creativity that make it valuable.

Key themes emerging from the summit:
-Collective intelligence may require friction and disagreement rather than perfect alignment
-Language shapes what AI can represent, preserve, and exclude
-Human knowledge should not be treated as anonymous training data
-Future benchmarks may need to measure agency, continual learning, and human flourishing, not just performance

“The deepest questions I heard in the AI research community were not only about technical progress, but about purpose: why are we building these systems, what role should they play in society, who does today's AI represent, and what responsibilities do we have to future generations?” said Ahmer Inam, co-founder of Cognisee PBC.“Today's AI still does not reflect the diversity of lives, cultures, languages, customs, and local realities across the world. ACI is our proposal for the missing governed collective intelligence layer - a framework for asking what the next era of AI should look like if it is truly meant to represent and serve humanity.”

The opening session, convened at Harvard by Ujjwal Kumar, co-founder of Cognisee PBC, framed ACI as a response to a structural gap in advanced AI: the absence of shared cognitive infrastructure for context, semantic coordination, consent, verification, revocability, and institutional accountability across human and AI agents.

“We have models that can pass bar exams, generate code, summarize research, and draft legislation. But if you ask who decided a model could learn from a community's knowledge, whether that consent can be verified, or whether it can be revoked, the answer is often silence,” said Kumar.“A person's knowledge is not raw material. It carries authorship, lineage, context, and decades of embodied judgment. If we treat it as training data, we have not built intelligence - we have built extraction.”

Participants represented organizations including Google, Harvard, MIT, Tufts University, NYU, McGill University, and other academic, research, and governance institutions.
Cognisee and participating collaborators are developing post-summit outputs, including a concise report and a jointly authored article outlining the ACI thesis. Planned workstreams will explore research roadmaps, evaluation frameworks, governance models, and pilot collaborations across clinical skill transfer, knowledge preservation, enterprise systems, physical AI, and regional intelligence infrastructure.

About Cognisee PBC
Cognisee is a frontier AI lab building computational models of cognition for high-stakes decisions and real-world action. The company works with nations, institutions, enterprises, and communities to develop sovereign intelligence systems that preserve critical knowledge, unlock trusted autonomy, and support human-AI collaboration across digital and physical workflows.

Learn more at .

MENAFN01062026003118003196ID1111193475