What You Need to Do

Read and respond to a workplace scenario that presents an ethical dilemma in at least two Discussion posts. In your response, address what you would do in the situation and suggest strategies to avoid the dilemma in the future. Discuss the posted options with your group. Use the step-by-step instructions below to complete the task.

How This Activity Connects to the Course

Ethical considerations matter in everything that you write in the workplace. You may well find yourself in a situation similar to this scenario once you graduate (or even sooner as an intern). This Discussion focuses on whether an AI system can accurately check the information in scientific publications.

Step-by-Step Instructions

Success Tip

You can use the online texts, any notes that you have, and the available course pages in Canvas for help as you work on these questions. You can also talk to one another.

  1. Read the scenario below and consider the ethical dilemma involved for Jamie.
  2. In your your first post, do the following:
    1. Present a position statement that answers the question “What would you do in this situation?”
      • Provide reasoning from your own personal code of ethics. Here you focus on what YOU think is right.
      • Provide reasoning from your discipline’s code of ethics, OR Provide at least one supporting statement or example from the Ethics Resource Module.
    2. Explain what Jamie’s and the Executive Committee should do to ensure their publications are accurate and reliable.
  3. Read the posts from others in your group, paying attention to how you agree or disagree on their solutions.
  4. Write at least one additional post to discuss the solution your group will recommend and the actions Jamie’s team should take to ensure their publications are accurate and reliable.

The Scenario

AI for Scientific Fact-Checking

A photo of a 27-year-old man in business casual dress, working at his desk on a laptop in a hotel room, preparing for the Executive Committee meeting in the morning.
Image created with Midjourney

ScyberPub, a prominent scientific publisher, is considering the integration of an AI fact-checker into its review process for incoming manuscripts. This AI tool would assess the credibility of sources cited in submissions, aiming to combat misinformation and enhance the accuracy of published research.

Scientists who publish frequently with ScyberPub are concerned about this potential automation. Determining the reliability of sources can be subjective and context-dependent. What one group considers a reliable source, another might see as biased.

Scenario

The Executive Committee of The World Biology Society is meeting to draft a statement on the use of AI in fact-checking and reviewing scientific publications. Jamie, a member of the Executive Committee, has volunteered to write the first draft of the standards for AI fact-checking and review. He must decide the capabilities that the AI fact-checker should have if it is to be trusted with verifying information in scientific publications. The committee has asked him to rank the following from Most Important (#1) to Least Important (#10).

Capabilities for Jamie to Review and Rank

  • Contextual Understanding: The AI should be capable of understanding the context in which information is presented, considering the nuances and specificities of scientific discourse.
  • Source Credibility Analysis: The AI should evaluate the credibility of sources by cross-referencing with established databases, considering factors such as the source’s history, peer reviews, and the reputation of the publishing institution.
  • Bias Detection: The AI should identify and flag potential biases in the sources or the information presented, providing a balanced view of the evidence.
  • Transparency: The AI should provide clear explanations for its decisions and assessments, allowing human reviewers to understand the rationale behind its conclusions.
  • Continuous Learning: The AI should be designed to continuously learn and update its knowledge base with new scientific information and evolving standards of credibility.
  • Cross-Disciplinary Verification: The AI should be able to verify information across multiple scientific disciplines, ensuring comprehensive fact-checking.
  • Historical Data Analysis: The AI should analyze historical data to recognize trends and validate current information against established scientific knowledge.
  • User Feedback Integration: The AI should incorporate feedback from users, including authors and reviewers, to improve its accuracy and reliability over time.
  • Ethical Compliance: The AI should adhere to ethical guidelines and standards set by scientific and professional organizations to ensure responsible use.
  • Scalability: The AI should be scalable to handle varying volumes of manuscripts without compromising the quality of fact-checking.

Issues to Consider as You Decide What Jamie Should Propose

  • Impact on Peer Review: How will the integration of AI affect the traditional peer review process and the roles of human reviewers?
  • Bias and Fairness: How can biases in the AI system be identified and mitigated to ensure fair evaluations of manuscripts?
  • Credibility Standards: What criteria should be used to determine the credibility of sources, and how should these criteria be developed and maintained?
  • Author Trust: How can the AI system be designed to maintain and build trust among authors who submit their work to ScyberPub?
  • Implementation Costs: What are the financial and logistical implications of implementing and maintaining an AI fact-checking system?

Conclusion

What standards should be established for the use of AI in fact-checking and reviewing scientific publications? Jamie must consider the potential impacts on the publishing process and the ethical requirements for reliability in science publications as he ranks the possibilities.