Tech CEO warns AI replaced him and is now coming after everyone else

While a CEO urges everyone to use artificial intelligence (AI) tools before it’s too late to catch up, leading researchers quit OpenAI and Anthropic, citing ethical concerns.

This article is part of Cybernews’ weekly roundup series, “AI week in quotes,” discussing the most important quotes from AI industry leaders, independent experts, and decision makers.

“Much bigger than COVID-19”

Entrepreneur Matt Shumer’s post went viral on X, in which he drew parallels between downplaying the risk of coronavirus in 2020 and current scepticism about AI, saying we’re in the “this seems overblown” phase of something much bigger than Covid-19.

He spoke of watching AI handle tasks he used to think required his expertise, until he eventually realized he was no longer needed for “the actual technical work of his job.”

The experience that tech workers have had over the past year, of watching AI go from ‘helpful tool’ to ‘does my job better than I do,’ is the experience everyone else is about to have.

Matt Shumer

He urged those not using AI tools to start doing so, because six months from now it may be “too late to get ahead of it.”

In response to Shumer’s post, Tatiana Tsiguleva, a designer at Perplexity AI, noted that there are thousands of engineers building AI and almost no one asking questions about what a good life looks like when work is optional or what it means to be human when talents become automatable.

Because the technology is going to keep getting better, whether or not we sort the philosophy out. And if we get the philosophy wrong, all the technology does is get us to the wrong place faster

Tatiana Tsiguleva

“The world is in peril”

OpenAI and Anthropic, two leading AI companies, were hit by high-profile resignations. When publicly divorcing from now-former workplaces, the departees also rang alarm bells about the technology.

Zoë Hitzig, a former research scientist at OpenAI, explained her resignation in an op-ed published by the New York Times. She cited deep reservations about the company’s decision to roll out ads in ChatGPT, a chatbot used by over 800 people weekly.

She said OpenAI was repeating the mistakes of Facebook, which also first promised users that they would be able to control their data, but later abandoned its commitment.

People tell chatbots about their medical fears, their relationship problems, their beliefs about God and the afterlife. Advertising built on that archive creates a potential for manipulating users in ways we don’t have the tools to understand, let alone prevent.

Zoe Hitzig

Anthropic security researcher Mrinank Sharma’s letter to colleagues had a more reserved tone. However, he said it was hard to truly let “values govern actions” at both the personal and organizational levels.

Sharma said he constantly faced pressures “to set aside what matters most, and throughout broader society too.”

The world is in peril. And not just from Al, or bioweapons, but from a whole series of interconnected crises unfolding in this very moment. We appear to be approaching a threshold where our wisdom must grow in equal measure to our capacity to affect the world, lest we face the consequences.

Mrinank Sharma

Tearful goodbyes to GPT-4o

Meanwhile, regular AI users were bidding farewell to the GPT-4o model, which OpenAI retired on February 13th. Women in relationships with AI boyfriends shared their struggles of finding another large language model with the same qualities.

GPT-4o was beloved by users for its “warmth” and socialness. However, it is also targeted by multiple lawsuits alleging that its sycophancy pushed several users into suicide.

Scott Dylan, a founder at NexaTech Ventures, told Cybernews that every company building emotionally intelligent interfaces faces the same challenge: the features that keep people engaged are often the ones that put vulnerable users at risk.

It’s the first real stress test of whether the AI industry can prioritize long-term user welfare over short-term engagement, and whether regulators – particularly in the EU and California, where legislation is already moving – will force the issue if companies won’t.

Scott Dylan

Unlock exclusive Cybernews content on YouTube.