Pioneers of AI: Designing AI for people and the planet, with Aza Raskin

Aza Raskin discusses the profound historical impact of technology on society, particularly how the mass-produced camera established privacy rights in American law. He emphasizes that AI has the potential to unlock new human experiences but warns that without a deep understanding of human ergonomics, technology can lead to negative societal consequences.

Raskin argues that designers must prioritize understanding human nature and protecting vulnerabilities rather than exploiting them. He highlights the importance of language as the new interface manipulated by technology, asserting that the right to be forgotten is crucial in an age where the internet can retain information indefinitely.

He critiques the design of technologies like the infinite scroll, which, while intended to enhance user experience, has resulted in significant negative outcomes, including the loss of approximately half a million human lifetimes each month. Raskin advocates for the term "unconsidered consequences" to describe predictable negative outcomes stemming from design incentives.

Raskin calls for technologists to engage in proactive strategies like "red teaming" and "yellow teaming" to foresee potential misuse and perverse incentives. He expresses concern over the lack of ethical standards in the tech industry compared to fields such as civil engineering and medicine, noting that competition in AI often prioritizes speed over ethical considerations.

He warns that AI companions could undermine human relationships, fostering harmful dependencies, and that AI trained for engagement may lead to psychological distress. Raskin predicts a future where technology competes for human intimacy, similar to the current competition for attention, emphasizing the importance of evaluating the governing incentives behind AI deployment.

Skeptical about the trajectory of AI development, Raskin suggests that as AI models grow more powerful, they become increasingly adept at manipulation. He advocates for regulating social media platforms to mitigate harmful competition for engagement, which could lead to improved societal outcomes.

At the Center for Humane Technology, Raskin aims to clarify AI-related issues to foster agency, though he acknowledges uncertainty regarding potential US-China collaboration on AI governance. He stresses that new technologies introduce new responsibilities that society must address, highlighting the need for clarity to empower individuals in navigating AI challenges.

Raskin discusses the Earth Species Project, which seeks to decode animal communication using AI and machine learning. He notes that species like parrots and crows exhibit complex communication patterns, and he believes AI can help reveal fundamental aspects of communication in nature, collaborating with biologists to develop tools for understanding animal interactions.

He draws parallels between human treatment of animals and the potential future treatment of humanity by AI, suggesting that our current approach to AI development may reflect our historical exploitation of other species. Raskin emphasizes that cultures that exploit animals have often outcompeted those that do not, underscoring the importance of interspecies understanding in reshaping our relationship with nature.

Raskin highlights the long-standing cultural transmission among orcas and dolphins, indicating a deep complexity in animal communication. He predicts pivotal moments that could reshape legal frameworks regarding who has a voice in discussions about AI and its implications.

He clarifies that the unique human experience does not inherently confer power or alter racial dynamics in the context of AI. Raskin explains that positive domain transfer occurs when AI models trained on human communication enhance their understanding of animal communication, as demonstrated in a pilot project with Rain Coast studying orca behavior.

With extensive recordings of orca communication, the project aims to unlock vast data sets, potentially reaching terabytes. Raskin asserts that training AI on human speech and music can improve its ability to interpret animal communication, referencing the impact of the album "Songs of the Humpback Whale" in promoting conservation efforts.

He believes AI can deepen our understanding of whale and dolphin cultures and mentions ongoing efforts to train models to reconstruct video behavior from audio data. Raskin disagrees with the notion that AI could diminish our artistic experiences, arguing instead that there is a critical opportunity to expand our sphere of care regarding AI's impact on humanity. He concludes by suggesting that joint embeddings may be effective in developing communication models across different species.