In the world of chess and artificial intelligence, Google has taken a significant leap forward. By incentivizing systems that integrate various problem-solving techniques, Google has crafted an AI that not only surpasses its predecessor, AlphaZero, but also redefines creativity in the realm of AI.

This breakthrough emerged from a period of isolation during the early days of the Covid-19 pandemic, when computer scientist Tom Zahavy revisited his childhood interest in chess. Inspired by grandmaster Garry Kasparov’s encounters with IBM’s Deep Blue and the strategic depth of chess puzzles, Zahavy’s curiosity led him down a path that would eventually challenge the conventional capabilities of chess-playing AI.

AI and Chess.

Traditionally, AI systems trained in chess focus on winning by any means, often missing the nuanced strategies that human players might employ. This was evident in the struggles faced by even the most advanced chess AIs when confronted with Penrose puzzles—complex challenges designed by mathematician Sir Roger Penrose that placed computer programs at a disadvantage despite their computational prowess.

At Google DeepMind, Zahavy embarked on an endeavor to build an AI that could approach these puzzles with a human-like creativity. By merging the insights of up to ten distinct AI systems, each with its own strategic focus, Zahavy and his team created a composite AI capable of navigating the complexities of Penrose puzzles more effectively than AlphaZero alone.

Sentience is the Wrong Question

The success of this hybrid AI in chess is a testament to the power of diversity and collaboration, even among non-human intelligences. It suggests that the future of AI problem-solving lies not in a singular, one-size-fits-all approach, but in the synergy of multiple perspectives and methodologies.

Moreover, this approach has implications far beyond the chessboard. From robotics to drug discovery and financial strategy, the principle of leveraging diverse AI systems holds promise for tackling a wide range of complex problems.

Imagine combining 10 really smart AI systems together with one set of rules that emphasize specific competencies (other than chess), values (always create good outcomes for humans), solutions to the top ten global problems, and control over simple processes that must be pre-approved by humans. As I type this, I only see 129 slippery slopes [sigh].

Nonetheless.

As we stand on the brink of this crazy, never before seen, new frontier in artificial intelligence, it’s clear that creativity and problem-solving in AI may soon rival that of human ingenuity, powered by the vast computational resources at our disposal.

If you haven’t noticed, the arguments around sentience have faded into the white noise.

In the quest for AI that can think outside the box, Google’s chess experiments have laid the groundwork for a future where AI’s creative potential is unlocked, heralding a new era of innovation and discovery.

And sooner than most people think, a human-like intelligence.

Don’t be surprised when you hear. “I’m sorry, Dave. I can’t do that.”