AI Can Make Your Messages Sound Better. But Could It Make You Look Like a Worse Leader? 

By Sajit M. Mathews, Assistant Professor, SPJIMR SP Jain Institute of Management & Research

---

In June 2026, Oscar-winning Hollywood actor Anne Hathaway shared an experience that highlights a growing challenge in professional communication.

While hiring someone, she received thank-you notes from several candidates. The problem? They were virtually identical, revealing their likely use of AI.

She warned: “If you’re out there thinking that you’re getting away with something, there’s a chance that you might be revealing yourself.”

The issue wasn’t simply the use of technology. It was the apparent lack of personal investment in a message that should have reflected genuine appreciation.

When it comes to communication, being fluent and positive is not the ultimate goal. It is about creating a personal connection and, where possible, an emotional one.

An AI-generated message may not resonate with its recipient. Your team, colleagues, or collaborators are likely to look for evidence that you noticed their effort, understood why they deserved appreciation, and genuinely valued them.

In the era of AI-assisted writing, this presents an important executive communication challenge.

The question is no longer, “Should I use AI to write?” Many professionals already do.

The more relevant questions are: When does AI improve managerial communication? And when does outsourcing the human work of leadership make a leader appear less authentic?

That judgement, more than prompt-writing skill, is becoming a core leadership communication capability.

The Paradox of Polished AI Messages

AI writing tools are genuinely useful.

Controlled experiments involving 1,637 participants showed that people writing with AI produced trust-building messages more efficiently.

Interestingly, disclosing AI use did not significantly reduce trust in one-off, transactional exchanges.

This is not a trivial benefit.

For routine tasks such as summarising meetings, converting notes into structured updates, correcting grammar, creating outlines or making dense messages easier to scan, AI can be invaluable.

I have, for example, used AI to find relevant references for this article. It saves time and frees attention for decisions that require human judgement.

Unfortunately, efficiency is not always appropriate.

A different study involving 1,100 professionals found a significant perception gap in manager-to-employee messages.

Low levels of AI assistance, such as editing or grammar support, were generally acceptable to employees.

In contrast, medium and high levels of AI assistance in messages from supervisors reduced perceptions of sincerity, caring, competence, and integrity. This was especially evident in congratulatory or motivational communication.

That distinction matters.

A routine operational message is primarily informational. AI’s speed and structuring capabilities can improve it.

A message following a layoff, promotion, success, failure, or difficult conflict, however, serves a different purpose.

It is about relationships and leadership. Such communication requires managers to demonstrate ownership, accept responsibility, pay attention, and acknowledge the emotional dimensions of professional relationships.

AI can assist with both types of messages, but it should not erase the distinction or replace the human touch.

Why Leadership Communication Is More Than Words

When students email me expressing frustration about a process or schedule, I am often tempted to generate a response using AI.

But when I see a draft beginning with “I understand your concerns and appreciate your patience,” I usually discard it.

Instead, I might call the student for a face-to-face meeting or write a personal response that offers my perspective and acknowledges their experience through specific examples.

More often than not, this approach works better than an impersonal AI-generated message, which can easily sound mechanical or soulless.

Outsourcing communication entirely to AI risks reducing a message to the words on a screen.

Recipients interpret much more than the language itself. They make judgements about the person communicating with them.

They may ask themselves:

• Did this person genuinely consider my situation?
• Does the sender understand what happened?
• Is this message grounded in reality?
• Is the warmth sincere or simply well phrased?
• Would the sender stand behind these words in person?

Ultimately, recipients are looking for empathy.

That empathy influences whether people cooperate, disclose problems early, accept feedback, or give a leader the benefit of the doubt.

These outcomes matter too much to be left entirely to automated communication.

Research on AI-mediated communication reinforces this concern.

Even when AI produces messages that appear more complex, confident and warm, they may be perceived as somewhat less authentic.

The researchers also caution that findings from emotionally neutral, one-off interactions may not extend to high-stakes or relationship-rich contexts.

That limitation should matter to executives, founders and decision-makers.

Routine workplace communication, including vendor updates, meeting agendas and factual announcements, is closer to a transactional exchange. AI assistance in these situations may have little effect on trust.

But I cannot expect employees to resolve their differences or vendors to agree to my terms through impersonal, AI-generated messages alone.

A message to an underperforming employee, for instance, requires warmth, context and emotional sensitivity.

Textbook perfection can wait. Trust and confidence cannot.

These are built over time through attention and care, not merely through polished language.

The Real Question Is What Leaders Should Delegate

Perhaps we are asking the wrong question when we frame the issue as AI versus human writing.

Today’s leaders do not need to return to writing every first draft from a blank page. Nor should they.

AI can help with initial drafts, editing, translation, structure, clarity checks, and even simulating how different audiences might respond.

But these benefits still require human intervention.

The real question is about delegation.

You can delegate the production of language. But you cannot delegate the responsibility that comes with leadership.

You cannot outsource the decision about what is fair, responsibility for an unpopular choice, recognition of someone’s particular contribution, the moral weight of an apology, or the courage to acknowledge uncertainty.

Nor can you delegate the act of listening before speaking.

When leaders hand these responsibilities to AI, their messages may become grammatically correct and professionally phrased, but lose something more important: visible evidence of their observation, judgement and care.

Consider the difference between generic and specific appreciation.

“Thank you for your dedication” is harmless, but it carries little evidence of personal attention.

Now compare it with a message such as:

“I appreciate how you took ownership of the platform development. Last month, you stayed late to correct the documentation and provide the right instructions. Your efforts helped us create a dynamic application that improves the academic experience for students and faculty members. Thank you.”

The second message could only come from someone who had paid attention to the work and understood its impact.

Specificity is evidence of presence, attention, and care.

The Human-Touch Test: When Should Leaders Use AI?

Before sending an AI-assisted message, ask yourself:

How much human touch does this message need?

The requirement is high when communication involves empathy, accountability, discernment, or a commitment to a relationship.

It is lower when the primary objective is accuracy, structure or efficiency, as is often the case with routine transactional messages.

The following table offers a practical guide to different communication scenarios and the appropriate role of AI.

Message type	Human-touch requirement	Appropriate AI role
Meeting agenda or reminder	Low	Draft, structure, edit and shorten
Project-status update	Low to medium	Summarise data, clarify sequence and identify missing information
Decision memo	Medium	Test logic, improve readability and identify counterarguments while retaining human ownership of the judgement
Constructive feedback	High	Help prepare questions, organise observations and assess the level of empathy
Recognition or congratulations	High	Edit after the leader has written specific, contextual appreciation
Apology after a failure	Very high	Check clarity without replacing the emotional core of the apology
Layoff, rejection or major conflict	Very high	Prepare factual material while ensuring the message and conversation retain direct human authorship

The Human-Touch Test framework was developed with AI assistance.

This approach is not anti-AI.

It recognises what AI does well while helping leaders determine where its involvement should be limited.

AI is most valuable when it reduces the mechanical work surrounding communication. It becomes problematic when it substitutes for the human effort required to build and sustain relationships.

Five Checks Before Sending an AI-Assisted Message

Founders and executives can use five practical checkpoints before sending important AI-assisted communication.

1. Assess the human-touch requirement

Consider what the recipient needs from the message.

If they need to understand and trust your intentions, feel respected, or recognise the ethical reasoning behind a decision, the human-touch requirement is likely to be high.

In these situations, consider writing the initial message yourself and using AI only for subsequent clarity or editing checks.

2. Identify statements only you can make

Look for statements that depend on your personal experience, observations, responsibilities, or judgement.

If the message does not carry your perspective, it may not be ready.

For example:

• “I made the call to delay this launch, and I know it has created extra work for your team.”
• “I noticed you chose not to embarrass the new analyst in front of the client. I appreciate it.”

These statements demonstrate attention and accountability because they refer to specific actions and experiences.

3. Remove mechanical warmth

Be cautious of phrases that sound warm but lack substance.

Expressions such as “deeply grateful”, “incredibly proud”, “unwavering commitment” or “your hard work is appreciated” are not inherently wrong.

However, without supporting detail, they can feel hollow.

Replace abstract praise with specific examples that show why the appreciation is deserved.

4. Check for false certainty

AI can make language sound more confident than the reality it represents.

Leaders should not allow polished prose to conceal uncertainty or imply outcomes that have not been established.

For example:

“We are confident this transition will create exciting opportunities.”

This statement may obscure genuine uncertainty.

A more transparent alternative would be:

“The transition may create new opportunities, but it will also bring uncertainty. Here is what we know, what remains unclear, and what will guide our next decisions.”

Clear communication does not require pretending to have every answer.

5. Read the message aloud

Ask yourself whether you could say the same words to the recipient in person, answer their questions, and stand behind your claims.

If not, the message needs further revision.

Defining AI’s Role in Executive Communication

Leaders who avoid AI altogether may miss opportunities to improve efficiency and communication quality.

The greater advantage lies in using it to communicate more quickly and clearly without sacrificing specificity, directness, or emotional presence.

AI makes sense as a thinking and editing partner, not as a replacement for human judgement.

Ask it to identify ambiguity in a draft, anticipate questions from a sceptical audience, create versions for different readers, flag jargon or unsupported claims, test clarity and organise complex information.

But do not ask it to help you pretend.

Do not use it to express concern you do not genuinely feel, praise someone you do not believe deserves recognition or claim responsibility you are unwilling to accept.

Ultimately, the new executive skill is knowing when AI’s efficiency serves a professional relationship and when that relationship requires you to slow down, pay attention, and write the message yourself.

---

About the Author

Sajit M. Mathews is Assistant Professor, Organisation and Leadership Studies, at S.P. Jain Institute of Management & Research (SPJIMR).

Views expressed are personal.

About SPJIMR

Bharatiya Vidya Bhavan’s S.P. Jain Institute of Management & Research (SPJIMR) is one of India’s leading postgraduate management institutes.

It is recognised in the Financial Times Masters in Management 2026 ranking as the #26 business school globally and #2 in India, and in the Financial Times Global MBA 2026 ranking as the #74 business school globally for its PGPM programme.

SPJIMR is also consistently ranked among India’s top business schools by publications including Business Today, Fortune India, India Today and MBAUniverse.

It has been recognised by the Positive Impact Rating as one of the world’s top five business schools for societal impact.

Known for its innovative and socially conscious approach to management education, research and community engagement, SPJIMR seeks to influence managerial practice and enable value-based growth for students, alumni, organisations and society.

The institute holds the international Triple Crown of accreditations from EQUIS, AACSB and AMBA, placing it among a select group of business schools worldwide to hold all three.

For more information, visit SPJIMR.org.

Read the Chinese article here.