AI Literacy Continuum

The stages are:

Not Yet Engaged → Uncritical Use → Informed Use → Critical Evaluation → Improvement

The stages are not meant to label people permanently or suggest that everyone moves through them in a perfectly linear way. A person's level of AI literacy can vary depending on the task, discipline, and context in which AI is being used.

Stage 1: Not Yet Engaged

At this stage, a person rarely or never uses generative AI.

Limited engagement may result from:

• Limited knowledge of AI tools

• Limited access

• Uncertainty about how AI works or how it should be used

• Mistrust of AI-generated information

• Ethical or privacy concerns

Not using AI is not necessarily a sign of low awareness. Understanding why someone does not engage with AI is an important part of assessing AI literacy.

Goal: Build enough understanding to make informed decisions about whether and when AI could be useful.

Stage 2: Uncritical Use

At this stage, a person uses generative AI but may rely on its outputs without consistently verifying or critically evaluating them.

Generative AI can produce responses that appear fluent, confident, and convincing even when the information is inaccurate or incomplete. This creates a risk of mistaking a plausible response for a reliable one.

For GES work, this could include accepting an AI-generated summary of a scientific paper, policy document, or biotechnology issue without checking the original source.

Goal: Develop habits of verification, source checking, and awareness of AI limitations.

Stage 3: Informed Use

Informed users understand the basic capabilities and limitations of generative AI and use it purposefully for tasks where it is appropriate.

This includes being able to:

• Choose appropriate tasks for AI assistance

• Understand that AI outputs require evaluation

• Verify important claims using credible sources

• Check whether references and citations are real

• Recognize that AI can produce inaccurate or incomplete information

• Protect information that should not be entered into public AI systems

• Apply human judgment to the final result

For GES, informed use means treating AI as an assistive tool, while maintaining responsibility for the accuracy and appropriateness of the work.

Goal: Use AI intentionally.

Stage 4: Critical Evaluation

Critical evaluation goes beyond checking whether an AI response is factually correct.

At this stage, users systematically examine AI outputs for:

• Accuracy and supporting evidence

• Bias or missing perspectives

• Privacy and confidentiality concerns

• Contextual appropriateness

• Oversimplification

• Uncertainty and limitations

• Potential consequences of errors

This is important for genetic engineering and society, where AI-generated information could influence how people understand scientific evidence, risks, benefits, regulation, or stakeholder perspectives.

Critical users ask "What might be missing, misleading, or inappropriate about this answer?"

Goal: Develop the judgment needed to evaluate AI outputs within a specific disciplinary and societal context.

Stage 5: Improvement

At the Improvement stage, users actively refine how they work with AI rather than simply accepting or rejecting its outputs.

This can include:

• Improving prompts and workflows

• Comparing different approaches or AI outputs

• Identifying and documenting limitations

• Developing evaluation or verification processes

• Improving the way AI is integrated into a larger workflow

• Helping others use generative AI responsibly

The focus changes from using AI effectively to improving the overall process through which humans and AI work together.

For example, a researcher might compare different prompting strategies for analyzing stakeholder discourse, establish a process for checking AI-generated findings against source material, document limitations, and refine the workflow for future use.

Goal: Develop more deliberate, transparent, and reliable human–AI workflows.