For GES work, responsible AI use should follow these principles:

1. Use AI for Appropriate Tasks

Before using generative AI, consider whether it is appropriate for the task.

AI can be useful for activities such as brainstorming, summarizing, literature synthesis, analyzing public or stakeholder discourse, drafting plain-language communication, coding, and data analysis. However, the appropriate level of AI involvement depends on the task, disciplinary context, and potential consequences of error.

Ask:

• Is AI appropriate for this task?

• What parts of the work require human expertise?

• What could go wrong if the AI output is incorrect?

2. Verify Important Information

Generative AI can produce confident-sounding information that is inaccurate or fabricated. Do not treat an AI response as authoritative just because it sounds convincing.

For important claims:

• Check information against credible external sources.

• Verify scientific, policy, and regulatory claims.

• Check that references and citations actually exist.

• Confirm that sources genuinely support the claims being made.

Verification is important when AI is used in biotechnology-related research, policy, or public communication.

3. Evaluate Evidence, Bias, and Context

Responsible use goes beyond asking whether an AI response is factually correct. Critical evaluation also considers what the response leaves out and whether its evidence is appropriate.

Consider:

• What assumptions does the AI output make?

• Is the evidence sufficient and representative?

• Are important perspectives missing?

• Could the output contain bias or oversimplify a complex issue?

• Does it accurately represent scientific uncertainty?

• Is the response appropriate for the disciplinary or societal context?

For GES work, this is particularly important when addressing contested scientific, ethical, policy, or community issues.

4. Protect Private and Sensitive Information

Do not enter private, confidential, proprietary, or otherwise sensitive information into public AI systems unless you have confirmed that the tool and its settings are appropriate for that information.

Consider:

• What information am I sharing?

• Who could potentially access or retain it?

• Does the information contain confidential research, personal information, unpublished data, or sensitive stakeholder information?

• Are there institutional or project-specific requirements governing its use?

5. Maintain Human Oversight

AI should support, not replace, disciplinary expertise and human judgment.

For consequential biotechnology-related work, qualified people should review AI-generated outputs before they are relied upon, shared, or used to inform decisions. Human reviewers should be able to question the output, identify limitations, correct errors, and decide whether the AI-assisted result is appropriate for its intended use.

6. Be Transparent About AI Use

Document and disclose generative AI use when it materially contributes to research, teaching, or public communication.

Depending on the context, this may include recording:

• Which AI tool was used

• What the AI was used for

• How the output was evaluated or modified

• What human judgment or expertise was applied

Transparency helps others understand how an AI-assisted result was produced and makes responsible evaluation easier.

7. Improve the Workflow, Not Just the Prompt

Responsible AI use can involve improving the entire workflow rather than simply trying to get a better response.

More advanced users can:

• Revise and compare prompts or approaches

• Develop verification procedures

• Document recurring limitations or failure patterns

• Build discipline-specific workflows

• Create evaluation criteria for AI outputs

• Share effective practices with others

The goal is to develop deliberate, reliable human–AI workflows rather than simply increase AI use.

GES Approach

Choose → Verify → Evaluate → Protect → Review → Disclose → Improve

Responsible AI literacy is about knowing when, why, and how to use it responsibly in a particular research, communication, teaching, or engagement context. For GES, the emphasis is on moving beyond uncritical reliance toward informed use and critical evaluation, while recognizing that responsible decisions may sometimes mean choosing not to use AI.