Is the Artificial Intelligence Tool for Assessing Corruption Risks in Legislative Acts Already Operational?

The AI platform BestAgreement.ai was developed to detect and assess corruption risks in legislative acts. The AI model for this platform was trained on various types of information, including legislative texts, court rulings related to corruption cases, analytical reports from anti-corruption organizations, and academic studies highlighting problematic aspects of legislation. The model was also trained based on international anti-corruption standards and recommendations from organizations such as the UN, OECD, Transparency International, and others. This data has enabled the model to effectively identify potential corruption risks and provide recommendations for their mitigation.

After training the model, we decided to test how well the BestAgreement.ai platform could detect signs of corruption in draft laws from different countries. The process was straightforward. We found draft laws that had already been reviewed by experts for potential corruption risks associated with the implementation of these laws’ provisions. 

Next, we uploaded a specific draft law to the BestAgreement.ai platform and received an anti-corruption report in just 30 seconds, which identified potential corruption risks. 

We then compared the report provided by the BestAgreement.ai platform with the expert analysis of the draft law in terms of the completeness and logical consistency of the conclusions. 

The results of the BestAgreement.ai platform’s performance pleasantly surprised us. We did not expect that the AI tool would accurately identify most of the corruption factors. This tool can significantly save time in analyzing corruption risks in draft laws and serve as a reliable auxiliary tool for specialists in corruption risk assessment.

Below are examples of the analyzed draft laws and the results of comparing the reports generated by the AI on the BestAgreement.ai platform with traditional expert analyses of draft laws for the presence of corruption risks.

Test 1: Ukraine, the draft Law “On Veteran Entrepreneurship”

We decided to analyze the draft law “On Veterans’ Entrepreneurship,” submitted to the Verkhovna Rada of Ukraine in March 2024. The anti-corruption examination of this draft law was conducted by the independent analytical center “Institute of Legislative Ideas.” We translated the draft law into English and uploaded it to the BestAgreement.ai platform.

You can see how the platform operates in this video.

Video

As a result of the study, we prepared the following documents:

1. The draft of  “On Veteran Entrepreneurship”(translated, English)  – download;
2. The “traditional” legal expert analysis,  prepared  by the Institute of Legislative Ideas, Ukraine – download;
3. The AI analysis,  prepared by BestAgreement.ai –  download;
4. The analysis of differences between “The AI analysis” and “The “traditional” legal expert analysis” – download;

In the last document, you will find a detailed analysis and a comparative assessment of the corruption risk reports for the draft law “On Veterans’ Entrepreneurship” from the BestAgreement.ai platform and the Independent Analytical Center “Institute of Legislative Ideas.”

Summary:

1. Overall, the BestAgreement.ai platform performed significantly better at detecting corruption factors and risks in the draft law “On Veterans’ Entrepreneurship” compared to the expert opinion prepared by the Independent Analytical Center “Institute of Legislative Ideas.” Notably, the BestAgreement.ai platform specifically identifies potential corruption factors and risks, while the expert report focuses on describing problematic provisions of the draft law without specifying corruption factors.

2. The conclusions are 99% consistent between the results from the BestAgreement.ai platform and the expert opinion from the Independent Analytical Center “Institute of Legislative Ideas.” However, the results from the BestAgreement.ai platform were obtained in just 30 seconds, whereas the expert analysis of the same law takes over five days.

3. The BestAgreement.ai platform provides unbiased conclusions regarding corruption risks in legislative acts, in contrast to traditional expert analysis, which may be influenced by pressures from interested third parties to reach certain decisions.

Test 2: Moldova, Draft Law “On Piloting Voting by Mail”

Following the positive results from Test 1, we analyzed the draft law “On Piloting Voting by Mail” in Moldova. The report on the presence of corruption factors and risks in this draft law was prepared by the National Anti-Corruption Center of the Republic of Moldova. We translated this draft law into English and uploaded it to the BestAgreement.ai platform.

As a result of the conducted research, we prepared the following documents:

1. The draft law “The Law on Piloting Mail-in Voting” (the Republic of Moldova)  – download
2. The anti-corruption expertise report of the Law on Piloting Mail-in Voting prepared by the National Anti-Corruption Center of the Republic of Moldova – download
3. The AI analysis,  prepared by BestAgreement.ai –  download
4. Comparative Analysis of the anti-corruption expertise report of the Law on Piloting Mail-in Voting of the Republic of Moldova with the results of AI Analysis – download

In the last document, you can find a detailed analysis and a comparative review of the corruption risk reports on the draft law “The Law on Piloting Mail-in Voting” from the BestAgreement.ai platform and the National Anti-Corruption Center of the Republic of Moldova.

Summary:

1. BestAgreement.ai consistently identified 17 corruption factors and risks, whereas the traditional legislative analysis conducted by the National Anti-Corruption Center of the Republic of Moldova identified only 4 potential corruption factors.

2. The AI-based analysis provided detailed descriptions and explanations of potential issues, demonstrating its reliability and effectiveness in detecting legal conflicts and corruption factors in the draft law “On Piloting Voting by Mail.”

3. The results indicate that the BestAgreement.ai tool can complement traditional legal analyses by identifying potential issues that may go unnoticed by expert lawyers.

Conclusions from the Study Presented in the Article:

1. The BestAgreement.ai AI platform has demonstrated high efficiency in detecting corruption risks in legislative acts. A comparison with traditional analysis methods shows that BestAgreement.ai can perform the same work in 30 seconds, whereas experts require more than five days to assess a single legislative act. This makes the platform extremely valuable for analyzing legislative acts for corruption factors and risks.

2. The test results showed that the conclusions made by BestAgreement.ai align with traditional expert analysis by 99%. However, BestAgreement.ai provides more specific and detailed descriptions of corruption factors, allowing for better analysis of potential risks with the aim of addressing them during the legislative process.

3. Unlike traditional expert opinions, which may be influenced by interested parties, BestAgreement.ai ensures an impartial and objective analysis. This significantly enhances the reliability of the conclusions, which is particularly important in the field of anti-corruption efforts.

4. The use of BestAgreement.ai can complement and improve traditional legal analysis methods by detecting potential issues that may be overlooked by experts due to the limitations of human cognition in processing large amounts of information. This allows for a more comprehensive approach to assessing corruption risks in legislation.

5. The platform has proven its effectiveness with examples from Ukraine and Moldova, demonstrating its ability to work with legislative acts from different countries. This opens up possibilities for its application on an international level, helping governments and organizations in the fight against corruption on a global scale.