All IDCA News

By Loading

8 Dec 2022

Share on social media:

Meta Created a 'Human-Level' Artificial Intelligence to Play Board Games

Facebook CEO Mark ZuckerbergGetty Images

For those who have never played Diplomacy and wonder what the big deal is, it's a board game first released in the 1950s in which people sit around a table (or break off into rooms) and decide what to accomplish.

No cards or dice are involved in the game; humans communicating with each other determines everything.

So for the creators of an AI to claim that it is playing at a human level in a game like this is a pretty bold statement that would likely be backed up if it said the AI is functioning on two levels; one which computes the progress and status of the game, and another which communicates with the humans who are playing.

One of their star endorsements is Diplomacy World Champion Andrew Goff. He claims, "A lot of human players will soften their approach, or they'll start getting motivated by revenge, and CICERO never does that. It just plays the situation as it sees it. So it's ruthless in executing its strategy, but it's not ruthless in a way that annoys or frustrates other players."

While CICERO is playing well enough to keep up with humans, it could be better, as Goff says. Meta states in a blog post: CICERO "sometimes generates inconsistent dialogue that can undermine its objectives,"

To me, every example of the communication they provide, for example below, makes it appear as if they are a crazed office worker afraid that if they don't include exclamation points at the end of their sentences, then you will think they are not good.

There is no ultimate goal with this program except to use Diplomacy as a sandbox by which to advance human-AI interaction:

While CICERO can only play Diplomacy, the technology behind this achievement is relevant to many real-world applications. Controlling natural language generation via planning and RL could, for example, ease communication barriers between humans and AI-powered agents.

For instance, today's AI assistants excel at simple question-answering tasks, like telling you the weather, but what if they could maintain a long-term conversation to teach you a new skill?

Alternatively, imagine a video game in which the non-player characters (NPCs) could plan and converse like people do — understanding your motivations and adapting the conversation accordingly — to help you on your quest of storming the castle.

It's possible that I'm not a billionaire, high-ranking executive in the Facebook company, but that doesn't change the fact that nobody, other than AI specialists and the company's expenditures, cares about these assistants I have. Instead of spending time and money on research, we could hire actual people or just humans. I'd have more peace of mind instead of worrying all day that the computer knows too much. He added.

Follow us on social media: