Integrating Artificial Intelligence (AI) into software systems has significantly enhanced their capabilities while escalating energy demands. Ensemble learning, combining predictions from multiple models to form a single prediction, intensifies this problem due to cumulative energy consumption. This paper presents a novel approach to model selection that addresses the challenge of balancing the accuracy of AI models with their energy consumption in a live AI ensemble system. We explore how reducing the number of models or improving the efficiency of model usage within an ensemble during inference can reduce energy demands without substantially sacrificing accuracy. This study introduces and evaluates two model selection strategies, Static and Dynamic, for optimizing ensemble learning systems' performance while minimizing energy usage. Our results demonstrate that the Static strategy improves the F1 score beyond the baseline, reducing average energy usage from 100% from the full ensemble to 62%. The Dynamic strategy further enhances F1 scores, using on average 76% compared to 100% of the full ensemble. Moreover, we propose an approach that balances accuracy with resource consumption, significantly reducing energy usage without substantially impacting accuracy. This method decreased the average energy usage of the Static strategy from approximately 62% to 14%, and for the Dynamic strategy, from around 76% to 57%. Our field study of Green AI using an operational AI system developed by a large professional services provider shows the practical applicability of adopting energy-conscious model selection strategies in live production environments.

Test amplification generates new test cases that improve the coverage of an existing test suite. To convince developers to integrate these new test cases into their test suite, it is crucial to convey the behavior and the improvement in coverage that the amplified test case provides. In this paper, we present NATIC, an approach to generate names for amplified test cases based on the methods they additionally cover, compared to the existing test suite. In a survey among 16 participants with a background in Computer Science, we show that the test names generated by NATIC are valued similarly to names written by experts. According to the participants, the names generated by NATIC outperform expert-written names with respect to informing about coverage improvement, but lack in conveying a test's behavior. Finally, we discuss how a restriction to two mentioned methods per name would improve the understandability of the test names generated by NATIC.