In Sam Altman’s vision for future, we all buy cheap intelligence from him

OpenAI Sam Altman envisions a future where artificial intelligence (AI) is bought like electricity or water. Millions of Americans who struggle to pay their utility bills may not be too fascinated with his vision.

Altman’s remarks come as OpenAI is hit by a major resignation over its deal with the US military. Meanwhile lawmakers demand answers over AI’s role in a strike on an Iranian school.

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

Intelligence “too cheap to meter”

OpenAI CEO Sam Altman, in an appearance at BlackRock’s US Infrastructure Summit on Wednesday, laid out his vision for the future, where “intelligence is a utility, like electricity or water, and people buy it from us on a meter.”

One of the most important things in the future is that we make intelligence, to borrow an old phrase from the energy industry that didn’t quite work: ‘Too cheap to meter.’

Sam Altman

Altman’s remarks may sound ironic, given that data centers necessary for AI generate massive electricity demand, and the costs of expanding infrastructure are often passed on to neighboring households.

Moreover, about one in six households in the US are behind on their energy bills, according to the National Energy Assistance Directors Association.

John Nosta, author and the founder of NOSTALAB, wrote that the phrase seems to capture “something larger than a commercial pricing model.”

It suggested a shift in how we may soon begin to understand intelligence itself less as a human capacity to be cultivated than as an external service to be accessed.

John Nosta

AI shouldn’t make life-or-death decisions

OpenAI has faced intense criticism after it announced the deal with the Pentagon in late February. The company replaced its competitor, Anthropic, which was blacklisted after objecting to its tools being used for domestic mass surveillance and autonomous weapons.

As calls to boycott its chatbot, ChatGPT, grew, OpenAI robotics lead Caitlin Kalinowski announced her resignation from the company. She said AI has “an important role in national security,” but said there are red lines for its use.

But surveillance of Americans without judicial oversight and lethal autonomy without human authorization are lines that deserved more deliberation than they got. This was about principle, not people.

Caitlin Kalinowski

At the same time, there’s growing suspicion that Anthropic’s Claude was involved in the US military attack on a school in Iran, leaving at least 175 people dead, most of them children.

Bo Young Lee, CEO of AI4ALL, wrote on LinkedIn that large language models (LLMs) were never designed to make “life or death decisions.”

If there was ever an argument that AI should not be used for any situation where error matters, the names and faces of those 175 children should be that argument.

Bo Young Lee

Two sources confirmed to NBC News that the military uses Palantir’s software, which relies in part on Anthropic’s Claude AI systems, to identify potential targets in the ongoing attacks.

However, the investigation is still ongoing, and the role of AI systems in the attack cannot be confirmed.

“Stop the monster AI”

Howard Yu, an author and professor at IMD Business School in Switzerland, compared AI approaches across Asia and Europe in a LinkedIn post.

He explained that people in Asia are used to “creative destruction” and move up the value chain as long as they learn. Meanwhile, in Europe, AI reignites the instinct to “protect what exists,” especially when amplified by apocalyptic talk of general AI.

In Shenzhen, people hear ‘AI’ and say, ‘Let's go.’ In Germany, the reaction is ‘Stop this monster.’ I travel between East and West regularly, and this gap in mindset is becoming one of the biggest predictors of who pulls ahead.

Howard Yu

Should AI replace journalists?

Burt Herman, a journalist and founder of Hacks/Hackers, shared his Slack message he sent to the Associated Press staffers as part of a discussion about AI in the newsroom.

He said that while he supports human work, there are tasks AI can do a pretty good job at, such as synthesizing reports, adding background, or rewriting for different formats.

Herman argued that AI quickly writing 500 words with context and background details, and adapting to every reading level, “could actually be better for the audience versus not receiving the information at all.”

Reporters who understand AI’s strengths and weaknesses will be highly valued in their organizations, knowing what only they can do to serve their community and what they can delegate.

Burt Herman

Unlock more exclusive Cybernews content on YouTube.