Sam Altman warns people are beginning to speak like AI

Social media feels less human

Sam Altman, the CEO of OpenAI, says social media posts today often feel strangely fake compared to just a few years ago. He noticed this while scrolling through discussions about Codex, the company’s new tool for developers.

Even though he knew the excitement around Codex was genuine, the posts themselves gave him the impression of being written by bots.

His reaction highlights how hard it has become to tell the difference between real human voices and automated accounts in online spaces.

People pick up AI quirks

Altman pointed out that one reason for this change may be people unconsciously imitating the way large language models write. Words and phrases once tied to chatbots are now slipping into everyday posts.

He believes highly active internet users move in similar directions, creating language patterns that look remarkably alike.

This blending of human habits with machine phrasing is part of why social platforms now feel less authentic and more robotic than they used to.

The hype swings too wildly

Another factor Altman mentioned is the extreme swings in online conversations about AI tools. One moment, users claim something is finished or outdated, and the next moment, they celebrate it as groundbreaking.

This back and forth creates a kind of artificial intensity that makes the whole conversation feel unnatural. Instead of balanced discussions, platforms often amplify exaggerated highs and lows, leaving many posts sounding like they were designed only to stir up engagement.

Platforms reward engagement

Altman also noted that social media companies optimize their platforms to keep people clicking. When engagement drives revenue, posts often follow a certain style meant to grab attention quickly.

This pressure can lead to writing that feels more like AI generated text than genuine human expression. The monetization system itself shapes how creators write, further blurring the line between authentic voices and content designed only to perform well online.

Paul Graham sees it too

Paul Graham, cofounder of Y Combinator, agreed with Altman’s observations. He said he also noticed many posts online now resemble language created by bots.

According to Graham, this shift is not just from large groups running fake accounts. Individual users trying to gain influence are also turning to AI tools to make their content more appealing, adding to the mix of machine like voices online.

Substack CEO raises alarm

Chris Best, CEO of Substack, has also shared worries about what he calls AI slop. He predicts that low effort, AI produced content will fill digital platforms in the years ahead.

He explained that such posts are designed mainly to attract clicks rather than to inform or entertain. This flood of shallow material could crowd out meaningful conversations, making it harder for people to find genuine voices on the internet.

Studies show vocabulary shifts

ChatGPT has influenced how people write and speak, with millions of texts showing a noticeable rise in certain words the AI commonly uses.

Terms like delve, examine, and explore appeared more often in emails, essays, and even YouTube content. Within eighteen months of ChatGPT’s release, people had already started adopting this style, showing how quickly machine generated patterns can shape human language.

Standard English takes the lead

ChatGPT responses can unintentionally reinforce dialect discrimination, often favoring Standard American English in its output while undervaluing non-standard varieties.

This creates challenges for users with different dialects or regional speech styles, since the AI often smooths their voices into one uniform style. Over time, this influence could push people to adjust how they communicate online to match the tone of the technology.

A surgeon customizes ChatGPT

Vaikunthan Rajaratnam, a nerve surgeon, described how he worked to make ChatGPT sound like him. By refining prompts and testing responses, he adjusted the tool to reflect his own vocabulary and thinking.

He said this helped bring clarity and structure to his work but also came with tradeoffs. While the results sounded eerily familiar, he worried about losing authenticity and regional expression in the process.

Billionaire points to humility gap

Mark Cuban, former Shark Tank investor, said there is one thing humans still do better than AI. He explained that artificial intelligence struggles to admit when it does not know something.

Cuban argued that this shortcoming gives humans an advantage, since humility and honesty are valuable traits in conversations. His comment served as a reminder that no matter how advanced AI becomes, it cannot fully replicate human self awareness.

Bots dominate online traffic

Data from Imperva showed that more than half of all internet traffic in 2024 was non human. Much of this came from automated systems, including those powered by large language models.

This statistic underlines why platforms feel increasingly artificial. With so many automated accounts and tools shaping the flow of information, even authentic voices can be drowned out in a sea of machine generated chatter.

Reddit users joke about Codex

Altman’s concern began after reading a flood of Reddit posts praising Codex. The posts sounded so similar that one user joked about whether switching to Codex required announcing it online.

For Altman, this highlighted how difficult it is to separate genuine enthusiasm from scripted content. Even with real growth behind the tool, the surrounding conversations gave the impression of being orchestrated rather than organic.

Echo chambers form quickly

Researchers at the University of Amsterdam once created a social network made entirely of bots. Surprisingly, the bots soon formed their own cliques and echo chambers.

This experiment showed how even automated systems mimic human social behaviors online. If bots can build insular communities, it becomes even harder for real people to distinguish where authentic discussion ends and machine made chatter begins.

OpenAI faces astroturf claims

Altman admitted that his company has been targeted by astroturfing in the past. This means outside groups may have posted content disguised as genuine user opinions to push their own agendas.

Because of this history, he said he is extra cautious when reading overly positive or negative posts about OpenAI products. It becomes difficult to know whether the enthusiasm or criticism is truly authentic or manufactured to sway public opinion.

AI’s influence on everyday speech

Beyond social media, some people notice changes in daily conversations influenced by AI. A researcher from the Max Planck Institute said his own vocabulary shifted within a year of using ChatGPT.

Words and phrasing from the system slowly crept into his emails and academic writing. This suggests that AI is not only shaping what we see online but also the way people talk and think in their personal and professional lives.

Looking for the breakthrough that could redefine AI in daily life? Explore GPT-5 now speaks, and it could change how we interact with AI daily.

The future of human voices

Altman’s observations raise bigger questions about the future of communication. If bots and human users both adopt similar styles, the internet could lose the sense of individuality that once defined it.

At the same time, some people are finding ways to adapt AI tools to match their own voices. Whether this trend preserves authenticity or erases it remains uncertain, leaving society to decide how much machine influence is acceptable in human conversation.

It raises questions about which voices will dominate, and how OpenAI could soon be the world’s most valuable startup, but Sam Altman won’t be the richest man.

Do you think people are starting to sound more like AI online? Share your thoughts in the comments and let us know if you’ve noticed the change too.

Read More From This Brand:

• Sam Altman admits GPT-5 launch flop
• Sam Altman warns of looming AI fraud wave
• Sam Altman flags privacy risks in ChatGPT therapy

Don’t forget to follow us for more exclusive content right here on MSN.

If you like this story, you’ll LOVE our Free email newsletter. Join today and be the first to receive stories like these.

This slideshow was made with AI assistance and human editing.