Object 33

Negotiation Agents

2030, Unified Korea

This simple interface may at first appear to be a mimic script, but it represented a step-change in humanity's use of external computing power for everyday decision-making:

Imagine that there's a company — let's call it Alpha — that has invented a revolutionary type of vehicle, so good that it could stand to make a few people very, very rich. But mere days after their vehicle goes on sale, another company — Beta — launches a similar vehicle.

To Alpha's eyes, it's not just similar — it's a copy based on valuable confidential information. They believe that a senior employee who recently left their company may have stolen confidential information and sold it to Beta. Alpha sues Beta, and things unfold as they ever have in a lawsuit, with parties obtaining evidence from one another via the discovery process.

Now imagine that you are a lawyer representing Alpha. In response to your requests Beta have sent you:

523 megabytes of text

7,492 hours of audio recordings

4,830 hours of video

5.3 petabytes of miscellaneous data including 3D reverse-engineered models, mimic scripting agents, company server logs, and some advanced biometric data.

How do you sift through that asteroid of information and find the nuggets of evidence that could constitute a coherent case? In the early 21st century, it would have taken dozens if not hundreds of skilled workers many months just to read everything, let alone process and understand it. Even then, it's likely that they would miss important details.

Luckily, it's not 2010 — it's 2030. You have access to affordable AI agents that can automatically scan and meaningfully summarise content, and if you have the funds, you can employ an amplified team to analyse the results and generate some potential next steps.

And then what? You still need to argue the case. If you're tech-savvy, you'll want to opt for a court system that allows lawyers to take advantage of mimic scripting and amplified teams. But if you want to save money and time, you might try a negotiation agent first — an AI system that can help opposing parties come to a mutually satisfying agreement.

It’s useful to remember where negotiation agents first came from. During the 20s, scientists, lawyers, and engineers in Unified Korea were tasked by the government to help fix the country's cripplingly slow and byzantine legal system, which reunification had only made worse. To help simulate the consequences of various solutions, they improved on a negotiation system used during the ASEAN (Association of Southeast Asian Nations) carbon-credit deadlock crisis, turning it into an early version of a negotiation agent. Many attribute the success of Unified Korea’s legal reforms to the negotiation agents, which were soon adapted to work worldwide.

More recently, negotiation agents have been improved so they can process and weigh vast quantities of data in a way that even an amplified team cannot. The reason you should use a negotiation agent in a dispute, however, isn’t because they’re faster or cheaper than humans — it’s because you can trust them to be perfectly impartial precisely because they aren’t human. When everyone concerned knows that the agents are unbribable, implacable, and utterly fair, parties that can’t agree on anything else can still agree to use an accredited open-source negotiation agent.

Now, naturally your client, Alpha, may not favour the use of a negotiation agent; they may be old-fashioned and worry that it cannot capture the subtleties of their position, or that it might not fight hard enough on their behalf. They are wrong, of course. When they see the record of the negotiation agent you're recommending, they'll come around — especially when you point out it'll cost a quarter as much as employing a human firm.

Finally, this is where you come in! Amid all the talk of summarising agents, mimic scripts, and amplified teams, we still need plain old humans. Alpha is old-fashioned. They're not comfortable talking with the negotiation agent — but you are. It's your job to act as a friendly, biological interface between the two. After all, that's what humans are best at: talking with other humans.