Human authorship, judgment and voice
Artificial intelligence is rapidly becoming more capable of producing competent prose. Generative AI can compose grammatical sentences, coherent paragraphs, plausible arguments, useful summaries and serviceable first drafts with remarkable speed. For speechwriters and preachers, this development should be welcomed as a significant extension of the tools available for research, reflection, drafting and editing. It does not diminish the importance of human authorship. Rather, it changes where the most important human contribution to writing occurs.
The central question is no longer simply whether a human or machine produced a particular sequence of words. More important questions concern whether those words are true, necessary, appropriate and worth hearing. Speechwriters must ask whether particular words should be spoken by this person, on this occasion, to this audience. Preachers face an additional theological question: does the sermon faithfully interpret Scripture and serve the people entrusted to the preacher’s pastoral care? As competent prose becomes increasingly automated, human contribution shifts towards attention, interpretation, judgment, wisdom, presence and responsibility.
This development need not be approached defensively. AI can enlarge the intellectual and creative resources available to writers, expose neglected questions, test arguments, compare interpretations, generate alternatives and accelerate routine editorial work. The aim should therefore be neither resistance nor uncritical dependence, but confident and disciplined engagement. AI is most valuable when it extends human attention and judgment rather than replacing them.
Fluency and the abundance of words
Generative AI is particularly adept at producing prose that appears finished. Grammatical correctness, rhetorical polish and orderly structure can nevertheless give mediocre thinking an appearance of authority. AI-generated material may contain factual errors, unsupported generalisations, conventional interpretations, repetition or clichés. More subtly, it may produce writing in which nothing is obviously wrong but little is memorable. Individual sentences can be competent while the argument remains intellectually thin.
Speechwriters and preachers should therefore resist equating fluency with quality. The relevant question is not merely whether a passage sounds impressive but whether it says something worth hearing. The principal challenge may increasingly be fluent language that requires careful human judgment concerning truth, relevance, proportion and depth.
AI also changes the economics of communication. Audiences are increasingly surrounded by competently generated emails, reports, advertisements, social-media posts, summaries, speeches and commentary. When plausible prose becomes abundant, attention becomes correspondingly scarce. This creates a premium on brevity, originality, concreteness and genuine human presence. A useful editorial question is therefore not only, “Is this well written?” but also, “Why should anyone listen to this?”
AI in the history of writing technologies
The opportunities presented by AI belong within a longer history of technological change. Printing, typewriters, word processors, spellcheckers, search engines and digital databases have all altered the ways people write, research, edit and publish. None eliminated authorship. Instead, each redistributed effort, accelerated particular tasks and changed expectations about competent writing.
Generative AI represents a further development because it participates not only in the mechanical production or retrieval of language but also in its apparent intellectual production. It can propose arguments, structures, interpretations and formulations. This makes it more consequential than many earlier tools, but not alien to the collaborative character of writing. Writers have always relied upon editors, colleagues, traditions, reference works and inherited forms.
The appropriate response is therefore neither technological alarmism nor reluctance to engage. AI should be learned, tested and incorporated thoughtfully into professional practice. The important question is not whether writers will use such tools, but how they can use them to increase intellectual range, improve communication and deepen responsibility.
Authorship, collaboration and agency
The familiar image of authorship as the work of an isolated individual has always been incomplete. Political leaders employ speechwriters; clergy consult commentaries and colleagues; academics work with editors; organisational statements are collaboratively drafted. Writers draw upon traditions, quotations, conversations and inherited forms. Authorship has long involved mediation.
AI introduces another form of mediation. It is therefore useful to distinguish composition from agency. Composition concerns who or what generated particular words. Agency concerns who determined the purpose, selected the argument, evaluated the evidence, exercised judgment and accepted responsibility for the result.
This distinction allows substantial use of AI without abandoning meaningful authorship. A writer may legitimately employ AI to generate options, reshape sentences, suggest structures or draft sections while remaining intellectually responsible for the whole. The ethical problem arises not from assistance itself but from accepting material without adequate judgment, verification or ownership.
The key question is therefore not merely, “Who wrote this sentence?” It is also, “Who exercised meaningful intellectual and moral agency over the whole?”
Editing as intellectual judgment
As acceptable sentences become easier to produce, judgment becomes more valuable. Skilled writers must distinguish between what could be said and what ought to be said. Editing therefore involves much more than correcting grammar or improving style. It requires decisions about argument, evidence, emphasis, sequence, proportion and tone. It asks whether a claim is true, necessary, clear and adequately supported, whether the argument follows logically, and what has been omitted.
AI can strengthen this process. It can compare alternative structures, identify repetition, test logical transitions, propose shorter formulations and flag claims requiring support. Used in this way, it becomes an unusually responsive editorial assistant. The human editor remains responsible for deciding which suggestions improve the work and which should be rejected.
Good editing is therefore a form of thinking. One of the most useful questions remains simple: “If this sentence were deleted, would anything important be lost?” In an age of inexpensive verbal production, disciplined subtraction may become as important as composition.
Information, interpretation and wisdom
Responsible AI use benefits from distinguishing information, interpretation and wisdom. AI is highly capable of locating, organising and synthesising information. It can also propose interpretations, compare arguments, identify patterns and reveal inconsistencies. These capacities make it exceptionally useful for research, drafting and editorial work.
Wisdom involves something more: determining what matters and why. It requires moral discrimination, sensitivity to context, recognition of ambiguity and awareness of consequences. A speech may contain accurate information and persuasive reasoning while displaying poor judgment. A sermon may demonstrate extensive biblical knowledge without exhibiting theological or pastoral wisdom.
This distinction should encourage rather than restrict AI use. The more effectively AI supplies information and possible interpretations, the more human attention can be directed towards discernment and application. The tool is especially valuable when it expands the range of possibilities before the writer rather than prematurely determining the conclusion.
Purpose should nevertheless precede prose. Writers should know what they want audiences to understand and what response they hope to encourage. AI can then assist in finding the strongest means of reaching that goal.
Speechwriting, voice and rhetoric
A speech is not simply an essay read aloud. It is language composed for embodiment by a particular person before a particular audience. AI tends naturally towards polished written prose, whereas effective speechwriting often requires shorter sentences, rhythmic variation, deliberate repetition, strategic pauses and concrete images. AI can assist with these tasks, particularly when asked to adapt prose for oral delivery.
Every significant speech should nevertheless be read aloud during editing. A sentence that is difficult to speak will often be difficult to hear. AI can help identify complexity, but oral delivery remains an indispensable test.
Good speechwriting must also preserve the speaker’s distinctive voice. AI is increasingly capable of adapting to preferred vocabulary, sentence length, rhetorical habits and levels of formality. Used carefully, this capacity can help refine rather than flatten individual style. The aim is not to make every speaker sound alike, but to strengthen the recognisable voice of the person who will deliver the words.
Classical rhetoric remains useful here. Effective communication involves logos, pathos and ethos. AI can assist substantially with reasoning and emotional resonance, while ethos remains closely connected to the character, history and credibility of the speaker. Language can enhance credibility, but it cannot replace the relationships and integrity upon which credibility finally depends.
Authenticity and human presence
AI can generate plausible emotional language, anecdotes and expressions of vulnerability. This is potentially useful, particularly when a speaker struggles to articulate genuine experience. Yet the emotional content must remain authentic. Speechwriters should not make speakers appear to have felt, witnessed or experienced something they have not, and preachers should avoid manufactured intimacy.
The appropriate use of AI is therefore to help express genuine experience more clearly, not to invent an emotional history. It can assist a speaker in finding language for grief, gratitude, hope, uncertainty or conviction, while the substance of those emotions must belong to the speaker.
Human authorship involves more than producing sentences. Human beings experience, remember, encounter, believe, doubt, suffer, hope and choose. AI can help such experience find clearer expression, but the significance of the words comes from the human life behind them.
From authorship to practice
The first challenge posed by AI is therefore conceptual: how should speechwriters and preachers understand authorship, judgment, voice and responsibility when machines can generate increasingly fluent prose? The answer is not to defend a romantic idea of solitary authorship, nor to retreat from new technology. It is to recognise that human agency increasingly lies in determining purpose, evaluating alternatives, preserving authenticity and taking responsibility for the finished work.
That conclusion leads naturally to a second set of questions. How should AI be incorporated into the daily practices of preaching and speechwriting? What forms of collaboration strengthen interpretation, formation, originality and ethical responsibility, and what disciplines are needed to use these tools well? Those practical questions are the subject of Part Two.
Rev Dr Rod Benson is General Secretary of the NSW Ecumenical Council and a minister of the Uniting Church in Australia serving at North Rocks Community Church in Sydney.
To read Part 2 of this article, click here.
Image source: author
All Told 