FORE20003 Ethics and Law of Forensics - Week 29: Bias Study Notes

The session will cover the following main topics:
- Types of cognitive bias
- The potential impact of cognitive bias on crime and criminal cases
- Approaches to mitigate the impact of bias
- Use and impact of artificial intelligence in forensic science

MLO1: Demonstrate an informed awareness of key aspects of the criminal justice system, focusing on ethical issues in relation to associated case law.
MLO2: Clearly explain the laws on pre-trial procedures, police powers, criminal law, and evidence, including principles and guidelines associated with the Police and Criminal Evidence Act 1984.
MLO3: Discuss, apply, and evaluate the law in case study scenarios based on real cases, reviewing procedures followed in each case in relation to the final court ruling.
MLO4: Comment on the legal requirements associated with the presentation of forensic evidence, emphasizing continuity of evidence and disclosure, including the consequences of not following correct procedures.
MLO5: Demonstrate good oral and presentational skills alongside a working knowledge of standards of evidence required for prosecution.
MLO6: Critically evaluate new knowledge and ideas to develop an understanding of best practices.
MLO7: Understand the roles and responsibilities of forensic science practitioners and witnesses within the legal system.
MLO8: Research legal sources using a range of electronic databases and retrieve relevant information.
MLO9: Express yourself clearly, succinctly, and accurately.
MLO10: Effectively utilize legal terminology in an impartial and comprehensive manner.

Definition: Cognitive biases are unintentional reasoning errors that individuals systematically make. They are tendencies, inclinations, or dispositions that cause distortions in information processing, leading to inaccurate or suboptimal outcomes.
**Sources of Bias:
- Environment
- Culture
- Case-specific factors
- Human nature and experience
- Case evidence (irrelevant information)
- Reference materials
- Base rate expectations
- Organizational factors
- Training and motivation
- Cognitive architecture and brain processes**
- These factors can interfere with accurate observations and inferences in forensic decision-making (Dror, 2017).

Real Case Analysis: Observing existing practices while acknowledging that some variables cannot be controlled, establishing factors that affect outcomes and decisions beyond mere evidence.
Case Simulations:
- Conducted by staging mock cases in controlled environments, allowing researchers to control variables during analysis.
Testing Practitioners:
- Involves inserting 'fake' submissions into normal case workflows and controlling for various factors to assess for cognitive bias.

Referencing studies and publications that demonstrate cognitive bias in forensic science, for example:
- Nakhaeizadeh, Dror, and Morgan (2014)
- Sunde & Dror (2019)
- Dror & Charlton (2006)
- Mattijssen et al. (2020)
Research Examples for Further Reading:
- Rapid identification and databases usage by Crime Scene Investigators (CSI) (Gruijter et al., 2017)
- Expectancy effects on CSI work (Eden et al., 2016)

Unconscious Bias Training:
- Promotes awareness of biases that impact decision-making in the workplace (EHRC, 2018).
Promoting Alternatives:
- Encouraging examiners to consider alternative hypotheses based on experience with prior cases.
Certification of Objectivity:
- Guidelines for maintaining objectivity among experts, highlighted by Lidén, Thiblin & Dror (2023).
Blind Peer-Review:
- Suggested method to reduce confirmation bias by preventing seniority from affecting decisions in cases of disagreement among experts (Mattijssen et al., 2020).

Linear Sequential Unmasking Protocols:
- Limit information given to experts and control the timing for sharing information (Dror & Kukucka, 2021).
Contextual/Case-Related Information Removal:
- Ensuring that only relevant data is presented to forensic investigators, withholding unnecessary case information until they arrive on site.
Case Coordination:
- Appointing a case manager to oversee information removal from case files to prevent exposure to contextual bias.
Order of Examination Considerations:
- Examining items in a specified order, e.g. determining which evidential item (like a skeleton) should be analyzed first (Davidson, Nakhaeizadeh & Rando 2021).

Predictive Policing & Security:
- AI uses historical crime data for analysis which aids in:
- Resource allocation
- Ensuring consistency of responses
- Enhancing safety measures
- Applications include predicting crime hotspots and assessing the likelihood of re-offending (e.g., Durham’s Harm Assessment Matrix).
- Facial recognition implemented in live CCTV by departments such as the Met Police and South Wales.
Concerns Related to AI:
- Transparency Issues: Lack of clarity regarding the functioning of AI systems.
- Surveillance, Privacy, and Bias:
- Potential for racial bias leading to over-policing and discrimination.
- Concerns over misidentification and the accuracy of AI assessments based on ethnicity and gender.
- The Equality & Human Rights Commission's call for suspension and legal modifications in 2020 regarding the use of face recognition technologies under existing laws (Data Protection Act 2018).

Unconscious Bias: Present in all individuals; while it cannot be completely eradicated, individuals can strive to minimize its impact through awareness and techniques designed to mitigate its effects on decision-making.
Cognitive Bias Types and Sources: Recognized biases in law enforcement, crime scene work, and forensic science leading to notable miscarriages of justice, prompting exonerations in recent years. The need for new protocols to address biases in investigation methods and technology use is critical. Continuous monitoring is necessary to ensure objective interpretations of evidence.