by Yesterday, hundreds of Eckersley’s community of friends and colleagues filled the pews for an unusual memorial service at the church-like sanctuary of the Internet Archive in San Francisco — a symposium featuring a series of lectures designed not only to commemorate Eckersley as a person, but to take a tour of his life’s work . Facing a shrine to Eckersley at the back of the hall, filled with his writings, his beloved racing bike and some samples of his Victorian goth-punk wardrobe, Turan, Gallagher and 10 other speakers gave presentations on Eckersley’s long list of contributions: his years propelling Silicon Valley toward better privacy-preserving technologies, his co-founding of a groundbreaking project to encrypt the entire web, and his late-stage focus on improving the security and ethics of AI.
The event also served as a sort of soft start for AOI, the organization that will carry on Eckersley’s work now after his death. Eckersley envisioned the institute as an incubator and applied laboratory that would work with major AI labs to address the problem that Eckersley now believed was perhaps even more important than the privacy and cybersecurity work he had been doing for decades had dedicated Career: The Future of Artificial Intelligence away from the forces that cause suffering in the world towards what he called “human thriving”.
“We need to make AI not just what we are, but what we want to be,” Turan said in his speech at the memorial after playing a recording of the phone call Eckersley used to recruit him. “So it can lift us in that direction.”
The mission that Eckersley conceived for AOI grew out of a growing awareness over the past decade that AI has an “alignment problem”: that its evolution is progressing at an ever-increasing rate, but with simplistic goals that are at odds with those of health and happiness of humanity. Rather than ushering in a paradise of abundance and creative leisure for all, Eckersley believed that AI on its current path is more likely to amplify all the forces already destroying the world: environmental degradation, exploitation of the poor, and rampant nationalism, to name a few .
The goal of AOI, as described by Turan and Gallagher, is not to try to limit the progress of AI, but to control it Goals away from those single-minded, destructive forces. They argue that this is humanity’s best hope to prevent, for example, hyper-intelligent software that can brainwash people through advertising or propaganda, corporations with god-like strategies and powers to harvest the last hydrocarbon from the earth, or automated hacking -Systems capable of invading any network in the world to cause global chaos. “AI failures will not look like nanobots suddenly crawling over us,” says Turan. “These are economic and environmental disasters that are going to look very recognizable, similar to what’s happening right now.”
Gallagher, now executive director of AOI, emphasizes that Eckersley’s vision for the institute was not that of an ominous Cassandra, but that of a shepherd who could lead AI to its idealistic dreams for the future. “He never thought about how to prevent a dystopia. His perpetually optimistic mindset was, ‘How do we make utopia?’” she says. “What can we do to build a better world and how can artificial intelligence contribute to the prosperity of mankind?”
