Facebook Claims To Have Developed A ‘Human-Level’ AI Board Game | TECHOCOTE

Facebook Claims To Have Developed A ‘Human-Level’ AI Board Game | TECHOCOTE

Facebook, or as we’re supposed to call them now, Meta, revealed earlier today that its CICERO artificial intelligence had reached “human-level performance” in the board game Diplomacy, which is remarkable because it’s a game based on human interaction rather than movements and man oeuvres like chess.

If you’ve never played Diplomacy and are wondering what the big deal is, it’s a board game first introduced in the 1950s in which individuals sit around a table (or break off into rooms) and negotiate.

Nothing is determined by dice or cards; everything is determined by humans conversing with other humans.

So for the makers of an AI to claim that it is playing at a “human level” in a game like this is rather a big assertion! One that Meta backs up by claiming that CICERO is actually functioning on two levels, one calculating game progress and status and the other attempting to communicate with human levels in a way we can comprehend and interact with.

Meta has enlisted the help of “Diplomacy World Champion” Andrew Goff to back up their accusations. “A lot of human players may temper their approach or get driven by retribution,” he argues, and CICERO never does. It just portrays the issue as it sees it. So it’s merciless in carrying out its plan, but not in a way that irritates or frustrates other players.”

That sounds ideal, but as Goff points out, it may be too ideal. This shows that, while CICERO is capable of keeping up with people, it is far from flawless. CICERO, as Meta says in a blog post, “sometimes generates inconsistent dialogue that can undermine its objectives,” and my criticism is that every example of its communication (like the one below) makes it look like a psychopathic office worker terrified that if they don’t end every sentence with “!!!” you’ll think they’re a terrible person.

The ultimate purpose of this programme, of course, is not to win board games. Diplomacy is just being used as a “sandbox” for “advancing human-AI interaction”:

While CICERO can only play Diplomacy, the technology that enabled this feat has numerous real-world implications. Controlling natural language creation by planning and RL, for example, might help to bridge communication gaps between humans and AI-powered bots.

Today’s AI assistants, for example, excel at simple question-answering tasks such as telling you the weather, but what if they could hold a long-term conversation with the goal of teaching you a new skill? Consider a video game in which non-player characters (NPCs) plan and converse like people, understanding your motivations and adapting the conversation accordingly, to assist you in your quest to storm the castle.

I’m not a billionaire Facebook executive, but instead of wasting time and money improving AI assistants, which nobody outside of AI research and company expenditure seems to care about, couldn’t we just…hire humans I can talk to instead?