The Likeability-Success Tradeoff: Results of the 2^nd Annual Human-Agent Automated Negotiating Agents Competition

Abstract

We present the results of the 2 nd Annual Human-Agent League of the Automated Negotiating Agent Competition. Building on the success of the previous year's results, a new challenge was issued that focused exploring the likeability-success tradeoff in negotiations. By examining a series of repeated negotiations, actions may affect the relationship between automated negotiating agents and their human competitors over time. The results presented herein support a more complex view of human-agent negotiation and capture of integrative potential (win-win solutions). We show that, although likeability is generally seen as a tradeoff to winning, agents are able to remain well-liked while winning if integrative potential is not discovered in a given negotiation. The results indicate that the top-performing agent in this competition took advantage of this loophole by engaging in favor exchange across negotiations (cross-game logrolling). These exploratory results provide information about the effects of different submitted "black-box" agents in humanagent negotiation and provide a state-of-the-art benchmark for human-agent design.

Varied Magnitude Favor Exchange in Human-Agent Negotiation

“…The final negotiation is structurally equal for both sides, with the items generally worth few points. This structure is similar to that of other work [23], and also provides an incentive to readily accept the favor in round 1 because there is an obvious structural basis to do so.…”

“…Temporallyaware agents have benefits that are most clearly illustrated by examining the additional value that can be claimed across multiple integrative negotiations. These benefits have been shown in ANAC 2018 [23], and provide the mathematical motivation and explanation for why favor-exchanging strategies are effective and sought-after.…”

“…Other work has examined effective negotiation strategies in human-agent competitions [23], as well as the use of favors in simple ultimatum games [21]. However, we are aiming to examine favor exchange in a new and more complex task (multi-issue bargaining).…”

Challenges and Main Results of the Automated Negotiating Agents Competition (ANAC) 2019

“…The human-agent league focuses on the myriad social effects present in mixed human-agent interactions. Indeed, understanding how humans negotiate has been a key question in business and psychological literature for many years-it is a complex social task [19,25,32,33]. But as automated agents are used more frequently in real-life applications (e.g., chatbots), we should design agents that are capable of interacting with humans in social settings.…”

“…Results from the first and second human-agent leagues (see [25,26]) show that while certain strategies may be effective in the short term, there is a tradeoff between agent likeability and agent success. To further examine this, the structure of the repeated negotiations were changed.…”