How AI Is Redefining Executive Presence: Why Empathy and Authenticity Are Now Core Currency

Many leaders now run their communication through an AI tool before it reaches anyone. Emails read more cleanly, summaries arrive in seconds, and a speech comes back tighter than the first draft. If you lead a team, you have probably noticed the effect from both directions: your own messages have improved, and so have everyone else's.

That raises a fair question about executive presence. If polish is available to anyone with a prompt, what makes a senior leader credible in the room? Recent research points to a consistent answer. The qualities that are difficult to automate, and difficult to fake in front of colleagues, are empathy that people can feel and authenticity they can verify.

What is executive presence, and what has AI changed about it?

Executive presence is the mix of qualities that leads people to trust a leader's judgement and want to follow their lead. In the widely cited framework from Sylvia Ann Hewlett and the Center for Talent Innovation, it has three parts: gravitas, communication and appearance. AI has made one of those, polished written communication, much cheaper to produce. That moves attention toward the part of presence that tools do not supply.

Hewlett's framework comes from a national survey and was published in her 2014 book, Executive Presence, years before generative AI tools reached everyday use. About two-thirds of the senior executives surveyed rated gravitas as the core ingredient. In that research, gravitas covers confidence under pressure, decisiveness, integrity, emotional intelligence, reputation and vision. These are traits that people infer from how a leader behaves over time, not from how a single document reads.

One feature of presence matters more now than it did a decade ago. It is a perception held by other people. A leader cannot declare it, and a tool cannot hold it on their behalf. When AI changes what colleagues expect to see, it changes how presence is read, even if the underlying qualities stay the same.

Why does AI make polish a weaker signal of leadership?

When anyone can produce a clear, well-organised message in seconds, clarity no longer separates one leader from another. Readers look elsewhere for evidence of who they are dealing with. That is an inference rather than a direct finding, but it fits what studies show about how AI-assisted messages are received.

A 2025 study in the International Journal of Business Communication, by Peter Cardon and Anthony Coman, surveyed more than 1,000 professionals about emails described as written with low, medium or high levels of AI assistance. Respondents generally saw the AI-assisted messages as effective and professional. The picture changed when they judged the sender. In the University of Florida's summary of the findings, only 40% to 52% of employees viewed a supervisor as sincere when the message involved high AI assistance, compared with 83% for low assistance. Ratings of professionalism for supervisors' messages also fell, from 95% to a range of 69% to 73%.

The study also found a perception gap. Professionals judged their own AI use leniently but were more sceptical of the same level of assistance from supervisors. Coman described the result as a tension between how a message is perceived and how its sender is perceived. For a leader, that distinction matters: the writing can pass inspection while confidence in the person behind it weakens.

These are survey responses to described scenarios, so they measure perceptions, not what happens inside any particular organisation. They still show a pattern worth taking seriously. Polish is easy to credit and easy to discount at the same time.

Can AI show empathy, and does that change how leaders are heard?

AI can produce messages that people experience as empathic, at least in controlled studies. The complication is that the effect weakens once people know a machine wrote the message. For leaders, this suggests that empathy depends on what the listener believes about the attention behind the words, as well as on the words themselves.

A 2024 paper in the Proceedings of the National Academy of Sciences, by Yidan Yin, Nan Jia and Cheryl Wakslak, tested this directly. AI-generated responses made recipients feel more heard than responses written by human participants, and the AI was better at detecting emotions. The human comparison group was untrained, so the result should not be read as AI outperforming skilled listeners. When recipients realised a message came from AI, they felt less heard. Third-party raters also found that the AI offered emotional support with more discipline and avoided excessive practical suggestions, which the authors note may be less effective at making people feel heard.

Two points follow for senior leaders. First, the finding about practical suggestions describes a familiar habit: moving quickly to solutions when someone describes a problem. Listening before advising is a skill, and it is one that people can watch you exercise. Second, the label effect fits an explanation the authors raise themselves. People may doubt that AI can truly understand or value them, and so discount what it says.

What happens to a leader's reputation when AI use is visible?

It can carry a social cost, although the size of the cost depends on context. In four preregistered experiments reported in the Proceedings of the National Academy of Sciences, people who used AI at work were judged by observers as less competent and less motivated, and those judgements affected how hypothetical job candidates were assessed. The result was not universal, which makes blanket rules in either direction hard to justify.

The authors, Jessica Reif, Richard Larrick and Jack Soll of Duke University, studied more than 4,400 participants. People expected to be judged negatively for using AI, and the experiments found that the expectation was justified. According to Duke's Fuqua School of Business, the penalty held regardless of the age, gender or occupation of the person described, and it disappeared when the evaluating manager used AI too.

The experiments used described workers and candidates rather than real colleagues, so they show how people judge in a controlled setting. A real leader's situation includes track record, relationships and organisational norms, all of which can change how AI use is read. What the research does show is that the risk is real enough to take into account. Competence and motivation are central to how senior leaders are judged, which is why this finding belongs in any discussion of presence.

Why does authenticity carry more weight when anyone can sound polished?

Authenticity is the part of trust that depends on people believing they are dealing with the real person, and a tool cannot establish that on a leader's behalf. When polish is widely available, consistency between what a leader says and how they behave becomes the more reliable signal. Authenticity does not mean disclosing everything. It means being recognisably the same person in the prepared and unprepared moments.

Frances Frei and Anne Morriss describe trust as resting on three drivers in their Harvard Business Review article on trust: authenticity, logic and empathy. People tend to trust a leader when they believe they are dealing with the genuine person, when they have faith in the leader's judgement and competence, and when they believe the leader cares about them. The authors also observe that everyone has a trust wobble, meaning a driver where trust is most likely to break down.

On a reasonable reading of that framework, AI helps most with logic. It can gather information, structure an argument and anticipate questions. That makes sound reasoning easier to produce for everyone, and so less distinguishing. Authenticity and empathy are the two drivers where a tool's contribution is least visible and most likely to be discounted once it is known. That reading is an interpretation, not a claim made by the authors, who wrote their article before generative AI was in common use.

Live moments make the difference visible: an unscripted question in a board meeting, a disagreement with a peer, a hard conversation with a direct report. No drafting is possible in those moments. What others see is how the leader listens, responds and handles not knowing.

Where can leaders use AI without weakening their presence?

Use it for work that supports thinking and preparation, and keep personal, relationship-oriented communication in your own words. The Cardon and Coman findings point the same way: employees accepted AI help with routine, informational messages, while heavy reliance on AI for messages such as team congratulations or motivational notes lowered perceived sincerity.

The sorting below is a practical suggestion based on that evidence, not a rule drawn from the studies.

Uses that tend to keep a leader's judgement and voice visible:

• Summarising background material before a meeting
• Testing an argument by asking for the strongest objections to it
• Producing a rough structure that the leader then rewrites in their own words
• Checking a draft for clarity or tone
• Preparing likely questions from a board or investors

Moments that benefit from the leader's own words:

• Acknowledging a difficult quarter or a setback
• Thanking an individual for specific work
• Giving feedback, especially critical feedback
• Apologising
• Responding to a team member in distress
• Explaining a decision that affects people's roles

Proportion appears to matter. In the survey, sincerity ratings were highest for low-assistance messages and fell as assistance rose to medium and high levels. Whether to disclose AI use is a separate question that the evidence does not settle. The studies cited here document how people react when AI use is known. They do not establish a best practice for disclosure.

How can a leader build empathy and authenticity into their presence?

Both are built through repeated, observable behaviour rather than through any single message or technique. Practical starting points are giving people undivided attention, being specific about what you know and do not know, and checking how you come across to others.

Attention comes first. In conversations that matter, put the phone away, let the other person finish, and say back what you heard before offering a view. This addresses the habit of jumping to practical suggestions, and it is something a drafted message cannot do for you.

Specificity signals attention. Naming the particular piece of work, the particular concern or the particular risk shows that you were present for it. Generic praise reads the same whether a person or a tool wrote it.

Candour about limits supports authenticity. Saying what you do not yet know, or that a decision remains open, shows reasoning that other people can inspect. It also matches the authenticity driver in the Frei and Morriss framework, which is about whether people feel they are dealing with the genuine person.

Follow-through makes words credible. Compare last quarter's commitments with what happened. Authenticity is judged over time, and a record of kept commitments is difficult for any tool to improve.

Outside feedback closes the perception gap. The Cardon and Coman survey found that professionals judge their own AI use more leniently than they judge their supervisors' use, which illustrates a wider problem: how a leader sees their communication and how others experience it can diverge. The EQ-i 2.0, which measures emotional intelligence, and 360-degree feedback, which collects other people's views, are two commonly used ways to compare self-perception with others' experience. Leaders who use psychometric assessments for this purpose sometimes pair them with coaching on executive presence, so that the feedback turns into specific changes in behaviour.

What does this mean for executive presence from here?

Nobody can say with confidence how perceptions of AI-assisted communication will settle, but current evidence points in one direction. Polish is becoming easier to obtain and less informative, while visible care and consistency carry more of the weight. Leaders do not need to avoid AI to protect their presence. They need to decide where their own voice and attention have to remain in view.

The evidence is also early. Most of it comes from surveys and experiments using described scenarios, not from long-term observation of real leadership teams. Attitudes toward AI are still changing, so the size of these effects may change too.

A simple test can help in the meantime. Before sending a message that matters, ask whether the recipient would feel differently if they saw exactly how it was made. If the answer is yes, the message probably needs more of your own voice.