According to the chapter, what is the difference between co-intelligence, collaborative intelligence, and collective intelligence in the context of human-AI collaboration?

Co-intelligence is a synergistic relationship where one human and one AI contribute complementary strengths with shared agency, with the human guiding and the AI suggesting or generating materials. Collaborative intelligence treats AI as a team member or collaborator with a defined role, relying on a division of labor that leverages each side's strengths. Collective intelligence works at a larger scale, involving groups of humans and multiple AI agents collaborating in networks to solve problems beyond the reach of any single person or machine.

In the source, co-intelligence describes a paired working relationship in which human and AI together produce results neither could easily achieve alone. The human keeps agency and provides context, judgment, and interpretation, while the AI contributes speed, pattern recognition, and content generation. A literature review where the AI summarizes sources and the human critiques and integrates them illustrates this mode. Collaborative intelligence is more team-like: the AI acts as a designated team member on a project with a clear division of labor. For example, AI might run numerous statistical regressions while human researchers interpret the results and design follow-up experiments. Humans still need to manage the AI's imperfections, much as they would a human colleague's limitations. Collective intelligence, traditionally the wisdom of crowds, becomes hybrid when groups of humans and AI agents work together in networks, each contributing knowledge or computations. An example is an online research community where scholars use various AI tools, and the aggregated, human-curated AI outputs yield discoveries no single person or machine could reach. This mode requires careful coordination and trust, with humans ultimately steering the collective effort. The source also cautions that these terms overlap and are used inconsistently in the literature, but together they express the idea of human-plus-AI augmented intelligence, with AI's autonomy ranging from a simple tool to an active teammate.