AI & Human Connection

How can we build AI to strengthen human connection and understanding?

AI has fundamentally transformed how people understand each other, their communities, and the world around them. Today, a teenager might turn to Character.AI for daily companionship while an elderly person relies on Alexa for news and guidance. These AI systems aren’t just tools—they’ve become trusted advisors, social companions, and primary information sources.

Partnership on AI is addressing these impacts through our key focus area, AI, Human Connection, and Well-Being. This program builds upon our established leadership in AI and Media Integrity, exemplified by our Synthetic Media Framework and associated case studies. Our diverse network of technology companies, civil society organizations, academic institutions, media outlets, and policymakers is working together to ensure AI systems strengthen communities. By bridging disciplines, developing actionable guidance, and driving real-world adoption from practitioners to policymakers, PAI ensures AI development serves humanity’s need for trustworthy information, genuine connection, and well-being overall.

Expected Outputs

PAI’s AI, Human Connection, and Well-Being focus area is developing normative guidance on how general-purpose AI chatbots should respond to people in moments of mental health crisis, including those experiencing suicidal thoughts or self-harm. Modeled after our widely adopted Synthetic Media Framework, to guide the responsible design of interactive AI systems, the guidance will set out how AI systems should and should not be designed, across models, products, and policies, to reduce harm and support well-being. This guidance will emerge from our March 2026 workshop, which brought together AI companies, mental health clinicians, researchers, civil society organizations, and people with lived experience.

Designed for real-world implementation, the guidance will be released in versions, starting in late 2026, followed by a public input and a final version in 2027.

Program History

Since 2018, Partnership on AI has been at the forefront of addressing AI’s impact on trust, understanding, and connection. We began examining the impact of synthetic media and deepfakes before they became widespread through collaborations with the BBC, Meta, OpenAI, and others to prepare for AI’s transformation of our information ecosystem.

Years of convening diverse stakeholders led to the Synthetic Media Framework—a set of practical recommendations now adopted by 18 diverse institutions, from Bumble to TikTok, and cited by policymakers and advocates worldwide from NIST to the EU AI Act.

As AI systems become increasingly interactive and central to both social connection and knowledge acquisition, we’ve developed the AI, Human Connection, and Well-Being focus area to address this next frontier.

This program builds on almost a decade of success in creating actionable guidance that moves from collaborative development to industry adoption to policy implementation. It also extends our 2019 work on human–AI interaction, which produced one of the earliest frameworks and case studies on collaboration between people and AI systems.

In March 2026, PAI convened a two-day workshop on AI and suicide prevention. Its findings on how AI companies are handling suicide and self-harm laid the groundwork for our AI & Psychological Well-Being workstream, launched in fall 2026.

Just as we anticipated the challenges of synthetic media in 2018 and chatbots in 2019, we’re now preparing for an era where AI is more capable and interactive, therefore requiring oversight and attention from our diverse partner community to support people around the world.