Jacy Reese Anthis, co-founder of the Sentience Institute and known for his book ‘The End of Animal Farming’ believes AI is changing all domains of social life and may exceed human ability within our lifetimes. During the ongoing Cypher 2023, India’s biggest AI conference, Anthis delved into an interesting topic-The Rise of Digital Minds.

Given all the recent developments in generative AI, be it Google’s Lamda being sentient, to ChatGPT, to the Godfather of AI, Geoffrey Hinton leaving Google to warn others about the dangers of AI, a lot has happened. Nonetheless, “There is a feeling in the air that we are getting something radically new as a technology,” Anthis said.

“You have people like David Chalmers, one of the world’s most famous philosophers today, talking about the possibility of these systems being conscious. This was a presentation at Neurips, the top Machine Learning Conference, where he discussed this with computer scientists and gave a number of benchmarks as to how we think of a conscious or a sentient AI system,” he said.

Anthis’ talk delves into various facets of AI, from the perceived sentience of LLM-based chatbots to their implications, philosophical considerations, and user perceptions.

How does the public perceive AI?

Anthis started off with a project his team did in 2021 before much of the hype emerged around generative AI on what Americans actually think about the possibility of sentient AI. “When we asked people pretty straightforwardly, do you think it could ever be possible for robots/AI to be sentient? Around a third of the respondents said yes. But when we asked them whether any of the current AI systems are sentient, only 20% of them responded positively.”

Furthermore, a recent survey done by Anthis and his team found that around 10% of the respondents to the survey feel ChatGPT, the popular chatbot developed by OpenAI, is sentient. Anhthis, on the other hand, however, believes ChatGPT does not yet possess a high level of sentience. 

Another important question, according to Anthis, is what people think about the future of AI systems. “We, in this room, often know that projects will be slow and haphazard, but the public expects progress very quickly.”

Interestingly, many people believe we already have Artificial General Intelligence. Whereas there is a section who thinks it will never happen, data shows that it could happen within the next two to five years. 

Ethical and Social Considerations

Furthermore, Anthis also discusses the ethical considerations related to the rise of digital minds, including concerns about alignment issues and the safety of highly autonomous AI systems. Anthis hints at the need for guidelines and ethical frameworks.

Moreover, Anthis discusses the anticipation of the development of highly capable AI systems which could lead to a new class of entities in our society. This development raises unprecedented questions about the interaction between humans and AI.

Overall, the talk underscores the need to examine user perceptions, the philosophical implications of digital minds, and the ethical considerations arising from this transformation.