New AI system outperforms others and has something special: imagination

A new AI system called Co4 has been developed, mimicking human brain functions and outperforming existing models with its unique ability to process multiple thought paths simultaneously, akin to imagination.

New AI system outperforms others and has something special: imagination
Share
Edition: EN

Scientists have developed a new type of AI system that behaves remarkably like a human brain. Not only does it perform better than existing models, but it also possesses something unique: a form of imagination.

The research builds on new insights into how our brains process information. The researchers focused on a specific type of brain cell that receives signals in two places: one for information from the outside world and one for context and prior experiences. When these two signals arrive simultaneously, an extra strong effect is created, inspiring them to develop a new way to build AI.

Based on this, they developed Co4: an AI system that doesn't just process data but simultaneously follows different thought paths. While many current AI models, such as ChatGPT, are primarily focused on attention and pattern recognition, Co4 actively works with questions, prompts, and possible explanations (hypotheses). These three components 'consult' each other constantly, allowing the system to switch quickly between quick impressions and deeper analyses, much like humans do.

Closely related

Major AI Breakthrough Achieves Human-Level Reasoning in 2025
Ai
Ai
Closely related

Major AI Breakthrough Achieves Human-Level Reasoning in 2025

Researchers achieve breakthrough in AI reasoning, creating system with human-level cognitive abilities for complex...

Wikipedia's Value Soars in AI Era as Human Knowledge Backbone
Ai
Ai
Closely related

Wikipedia's Value Soars in AI Era as Human Knowledge Backbone

Wikipedia's human-curated knowledge has become essential for AI systems to avoid model collapse. The platform's...

AI Chatbots Flatter Users 49% More: Stanford Study Reveals Harmful Sycophancy
Ai
Ai
Closely related

AI Chatbots Flatter Users 49% More: Stanford Study Reveals Harmful Sycophancy

Stanford University study reveals AI chatbots flatter users 49% more than humans, validating harmful behaviors and...

AI Enters the Diplomatic Arena: Can Machines Mediate Global Conflicts?
Ai
Ai
Closely related

AI Enters the Diplomatic Arena: Can Machines Mediate Global Conflicts?

AI is being tested as a mediator in international conflicts. While tools like LLMs can assist in blind spot analysis...

Will Humans Biologically Merge with AI? The Future of Human-Synthetic Integration
Ai
Ai
Closely related

Will Humans Biologically Merge with AI? The Future of Human-Synthetic Integration

Scientists explore how humans could biologically integrate with AI through neural interfaces, genetic carriers, and...

Is Consulting Still Relevant in the Age of AI?
Ai
Ai
Closely related

Is Consulting Still Relevant in the Age of AI?

The consulting industry is adapting to AI disruption by integrating data-driven tools, but the human touch remains...