OpenAI Is Hiring for a Place That Sounds Horrifying

“This will likely be a annoying job and also you’ll leap into the deep finish just about instantly,” OpenAI CEO Sam Altman wrote on X in his announcement of the “head of preparedness” job at OpenAI on Saturday.

In alternate for $555,000 per year, based on OpenAI’s job advert, the top of preparedness is meant to “develop, strengthen, and information,” the prevailing preparedness program inside OpenAI’s security techniques division. This aspect of OpenAI builds the safeguards that, in principle, make OpenAI’s fashions “behave as supposed in real-world settings.”

However hey, wait a minute, are they saying OpenAI’s fashions behave as supposed in real-world settings now? In 2025, ChatGPT continued to hallucinate in legal filings, attracted hundreds of FTC complaints, together with complaints that it was triggering psychological well being crises in customers, and evidently turned pictures of clothed women into bikini deepfakes. Sora needed to have its potential to make movies of figures like Martin Luther King, Jr. revoked as a result of customers had been abusing the privilege to make revered historic figures say basically anything.

When instances associated to issues with OpenAI merchandise attain the courts—as with the wrongful demise go well with filed by the household of Adam Raine, who, it’s alleged, acquired recommendation and encouragement from ChatGPT that led to his demise—there’s a authorized argument to be made that customers had been abusing OpenAI’s merchandise. In November, a submitting from OpenAI’s attorneys cited rule violations as a potential cause of Raine’s death.

Whether or not you purchase the abuse argument or not, it’s clearly a giant a part of the way in which OpenAI is smart of what its merchandise are doing in society. Altman acknowledges in his X publish concerning the head of preparedness job that the corporate’s fashions can affect individuals’s psychological well being, and might discover safety vulnerabilities. We’re, he says, “coming into a world the place we’d like extra nuanced understanding and measurement of how these capabilities could possibly be abused, and the way we are able to restrict these downsides each in our merchandise and on the planet, in a method that lets us all benefit from the great advantages.”

In spite of everything, if the objective had been purely to not ever trigger any hurt, the quickest method to ensure of that may be to simply take away ChatGPT and Sora from the market altogether.

The top of preparedness at OpenAI, then, is somebody who will thread this needle, and “[o]wn OpenAI’s preparedness technique end-to-end,” determining tips on how to consider the fashions for undesirable talents, and design methods to mitigate them. The advert says this individual must ”evolve the preparedness framework as new dangers, capabilities, or exterior expectations emerge.” This could solely imply determining new potential methods OpenAI merchandise may have the ability to hurt individuals or society, and give you the rubric for permitting the merchandise to exist, whereas demonstrating, presumably, that the dangers have been dulled sufficient that OpenAI isn’t legally chargeable for the seemingly inevitable future “downsides.” 

It could be unhealthy sufficient having to do all this for an organization that’s treading water, however OpenAI has to take drastic steps to usher in income and launch leading edge merchandise in a rush. In an interview final month, Altman strongly implied that he would take the corporate’s income from the place it’s now—apparently someplace north of $13 billion per yr—to $100 billion in lower than two years. Altman mentioned his firm’s “shopper system enterprise will likely be a big and essential factor,” and that “AI that may automate science will create enormous worth.”

So if you need to supervise “mitigation design” throughout new variations of OpenAI’s current merchandise, together with new bodily devices, and platforms that don’t exist but, however are imagined to do issues like “automate science,” all whereas the CEO is respiratory down your neck about needing to make roughly the same amount of annual revenue as Walt Disney the yr after subsequent, get pleasure from being the top of preparedness at OpenAI. Attempt to not fuck up all the world at your new job.