Splice CEO Sees a Conversation Cost in AI Emails

Splice CEO Kakul Srivastava says AI-generated emails are killing conversations, raising a question about whether faster writing improves workplace communication.

Writing an email faster is not the same as reaching an understanding faster. That distinction is at the heart of Splice CEO Kakul Srivastava’s view that artificial intelligence-generated emails are killing conversations.

Her criticism puts a practical question in front of businesses adopting AI writing tools: does the software help people communicate, or simply help them produce more polished messages? A smoother sentence can be useful. But if it obscures uncertainty, softens a necessary disagreement or adds text without adding meaning, the recipient may have more work to do.

Srivastava’s position was reported by The Verge. The verified facts establish her criticism, not a measured decline in communication quality or a company policy restricting AI. The distinction matters: this is an executive’s assessment of a technology’s effect on conversation, rather than evidence that every AI-assisted email fails.

The business test is understanding, not output

The attraction of AI-assisted email is straightforward. A user can ask software to draft a reply, organize a rough thought or adjust the tone of a message. Those functions may remove friction from writing, especially when someone knows what they want to say but struggles to express it.

The harder question is what happens after the message leaves the sender’s screen. An email is part of a shared task. Its value depends on whether someone else can understand the request, recognize the constraints and decide what to do next. Generating text is only part of that process.

Consider a hypothetical project update. A manager needs to explain that a deadline is uncertain because a decision remains unresolved. A polished draft might sound reassuring while failing to make that dependency explicit. The writing could look better even as the information becomes less useful. The problem would not be grammatical accuracy; it would be a mismatch between the message’s tone and the situation.

That example illustrates the concern without establishing how frequently it occurs. Businesses evaluating these tools should distinguish the convenience of producing a draft from the quality of the exchange it creates. Time saved by a sender can be offset by clarification work handed to a recipient.

Why a music platform makes this question relevant

Splice is a sample platform that producers use for one-shots and melodic loops. A one-shot is a discrete sound, while a melodic loop is a musical phrase that can repeat. Samples from the service have appeared in Lisa’s “Money” and Sabrina Carpenter’s “Espresso.”

Those uses place Splice within a creative workflow built around reusable material. Producers can start with sounds they did not record themselves and make choices about how those sounds fit a composition. The existence of a ready-made component does not settle what the finished work should communicate.

There is a useful parallel with assisted writing, though samples and generated emails are different products. In either case, access to material can make creation easier without removing the need for judgment. Selecting, editing and rejecting remain important. A tool’s usefulness depends partly on how deliberately a person uses it.

Srivastava’s criticism should not be stretched into a position on every creative use of AI. The supplied facts do not establish Splice’s AI strategy, a change to its service or a judgment about AI-generated music. Her stated concern is about emails and conversations. The platform’s role in music production provides context, not proof of a broader corporate stance.

Polished language can hide the point of an exchange

A productive workplace conversation often contains information that is awkward to state: a reservation, an incomplete idea, a disagreement or a request for help. These details can be more valuable than a message’s elegance. If a writing tool smooths them away, it may remove the very reason the exchange was necessary.

Imagine another hypothetical case: a colleague objects to a proposed approach but asks an AI tool to make the objection sound friendlier. That edit could help preserve a working relationship. It could also weaken the objection until the recipient mistakes resistance for agreement. The outcome depends on the instructions, the draft and the sender’s review.

This is not an argument that bluntness always beats courtesy. Tone matters, and assistance can help someone avoid accidental hostility. The relevant boundary is whether the edit preserves the sender’s meaning. A courteous disagreement still needs to be recognizable as a disagreement.

There is also a distinction between correcting expression and supplying substance. A tool that helps organize a decision already made plays a different role from one that fills an empty reply with plausible language. In the latter case, the sender must decide whether the generated message reflects an actual judgment they are prepared to defend.

What companies should watch in AI-assisted email

For employers, the practical response is to evaluate communication outcomes rather than celebrate draft production alone. Useful questions include whether recipients understand the requested action, whether important qualifications survive editing and whether exchanges resolve issues instead of extending them.

Review should focus on meaning as well as factual accuracy. Before sending an assisted draft, an employee can check whether it says what they intend, states what remains unknown and makes clear who needs to act. An accurate email can still be unhelpful if its purpose is buried.

Companies should also distinguish situations where assistance is valuable from situations where direct discussion may work better. Routine wording and difficult judgment are not interchangeable tasks. When the central issue is disagreement or uncertainty, another polished email may not be the best next step.

The question to watch is whether AI email tools make those distinctions easier for users to manage. Srivastava’s criticism directs attention to the part of productivity that a finished draft cannot demonstrate: whether people actually understand each other. For businesses, that is the more consequential test.

Key facts

• Executive: Kakul Srivastava is Splice’s CEO.
• AI criticism: Srivastava thinks AI-generated emails are killing conversations.
• Platform: Splice provides samples producers use for one-shots and melodic loops.
• Music uses: Splice samples have appeared in Lisa’s “Money” and Sabrina Carpenter’s “Espresso.”

Frequently asked questions

What does Kakul Srivastava think about AI emails?

Splice CEO Kakul Srivastava thinks AI-generated emails are killing conversations. Her criticism raises the distinction between producing polished text and helping people understand each other. The available facts do not establish a measured decline in communication quality, a ban on AI email tools or a broader rejection of artificial intelligence.

What is Splice?

Splice is a sample platform producers use for one-shots and melodic loops. One-shots are discrete sounds, while melodic loops are musical phrases that can repeat. Samples from the service have appeared in Lisa’s “Money” and Sabrina Carpenter’s “Espresso.” Kakul Srivastava is the company’s chief executive.

How could an AI-written email make communication worse?

An AI-written email could make communication worse if its polished wording obscures uncertainty, weakens a disagreement or buries the action a recipient needs to take. These are possible failure modes, not documented incidents in the supplied facts. The central test is whether the draft preserves the sender’s intended meaning.

Can AI assistance still be useful for workplace emails?

AI assistance can be useful when it helps a sender organize an existing thought, clarify wording or adjust tone without changing the substance. That potential benefit does not resolve Srivastava’s concern. A business still needs to assess whether recipients understand the message and whether the exchange helps them complete the underlying task.

Has Splice announced an AI email ban?

The supplied facts do not establish that Splice has announced an AI email ban or changed its workplace communication policy. They establish Srivastava’s role, her criticism of AI emails and the platform’s use in music production. Treating her opinion as a formal company restriction would go beyond the available information.

What should businesses check before sending AI-assisted emails?

Businesses can ask employees to check whether an AI-assisted draft preserves their meaning, identifies unresolved questions and makes the requested action clear. Reviewing grammar and factual accuracy alone may miss communication problems. The relevant outcome is whether the recipient understands what matters, not simply whether the sender produced the message quickly.

Sources

• Splice CEO Kakul Srivastava thinks AI emails are killing conversations — The Verge

Photo: markusspiske · Pixabay Content License — source