AI Algorithm in Judicial Justice
Overview of the State v. Loomis Case
Legal Context: The case highlights the use of proprietary algorithms in judicial decisions, particularly in assessing risk of recidivism.
Central Figure: Eric Loomis, whose fate was determined by an algorithm rather than human judgment.
Key Events in the Case
Court's Decision:
The court ruled that Loomis posed a high risk of recidivism, leading to denial of probation.
Loomis was sentenced to six years in prison based on an algorithmic decision, showing reliance on a proprietary software called COMPASS.
Role of Algorithm in Legal Decision-Making
COMPASS Software:
A proprietary tool used to assess the likelihood of a defendant committing further crimes.
Generated visual outputs (e.g., bar charts) which influenced judicial decisions.
Judge's Interaction with the Algorithm:
The judge relied on the risk score without significant engagement with Loomis’s defense.
Ethical and Legal Implications
AI in Legal Judgments:
The decision raises concerns about the use of AI in making life-altering legal judgments.
Loomis’ inability to scrutinize the algorithm’s decision-making process points to potential injustices.
Duties of Professionals Relying on AI:
Attorneys and HR professionals bear the responsibility of ensuring ethical and legal standards in decisions made using AI.
There is a warning for all professionals that their reliance on AI tools cannot replace due diligence and ethical considerations.
The Issue of Transparency in AI Tools
Right to Know - Due Process:
Loomis argued his due process rights were violated due to lack of transparency in how the risk score was computed.
The proprietary nature of COMPASS prevented him from verifying its scientific validity and assessing potential biases.
Trade Secrets vs. Individual Rights:
The court acknowledged the conflict between proprietary information and the rights of defendants to challenge evidentiary basis for judgments.
Court's Ruling and its Implications
Wisconsin Supreme Court Decision:
Although the court recognized the issues surrounding COMPASS, they upheld Loomis’ sentence.
The ruling allows judges to utilize secret risk assessments if accompanied by a significant warning.
Warning Label Established:
COMPASS scores are to be seen as supplementary evidence rather than definitive proof for sentencing.
Judges must inform that these scores should not be determinative in individual cases, emphasizing the need for a complete consideration of the individual's circumstance.
Group Data vs. Individual Reality:
The algorithms provide assessments based on group behavior statistics, which may not accurately reflect individual risk factors.
Call to Action for Professionals
Responsibilities of Legal and HR Professionals:
Professionals must act as a ‘human firewall’ ensuring that AI does not dictate judgments on liberty, employment, or reputations.
There is a strong emphasis on demanding transparency and accountability from AI systems used in decision-making.
Need for Education on AI Usage:
Subscribe to platforms like AI Justice Unpacked for further understanding of integrating AI responsibly in professional settings.