What is Open Source AI and why is this so controversial?

In this sharp and thought-provoking session from the Artificial Unintelligence Conference, Mo McElaney unpacks one of the most misunderstood and misused terms in technology today: “open source.”

McElaney introduces the concept of “open-washing” — when companies market their AI models as “open” while quietly withholding critical components like training data, weights, or full licenses. The result, she argues, is a dangerous illusion of transparency that undermines accountability and erodes public trust.

Through examples and clear ethical framing, McElaney challenges viewers to rethink what “open” should mean in an AI-driven world — not as a branding tool, but as a philosophy of stewardship, reproducibility, and shared responsibility.

Key ideas explored:
• What “open-washing” reveals about AI’s current transparency crisis
• Why partial openness can be more harmful than closed systems
• The need for clear, enforceable definitions of “open source” in AI
• How open source can return to its roots as a public good
• The social and ethical stakes of AI transparency