My application to an AI lab was rejected. I lacked pricing alignment.

Dear Candidate,

Thank you for your interest in joining Beneficence Frontier Systems, a fictional laboratory dedicated to ensuring that advanced intelligence benefits everyone able to complete our procurement process. After careful consideration, we will not be moving forward with your application for Head of Existential Strategy.

Your qualifications were strong. The hiring panel particularly appreciated your ambition, your willingness to speak in absolutes, and your complete lack of prior experience delivering what you promised. These are all valuable traits in a rapidly evolving field.

However, several concerns arose regarding your alignment with our pricing page.

You misunderstood our mission

When asked how you would ensure that humanity retained control over transformative technology, you proposed giving humanity a meaningful say in its governance. While inventive, this answer failed to reflect the current organizational structure.

Our mission refers to humanity as a beneficiary, not as an administrator. The distinction is clearly explained in the settings menu, which is available to Enterprise customers.

You also asked who would be accountable if a system made an expensive mistake. The correct answer is “we take these concerns seriously.” You instead requested a name. We consider this an unnecessarily confrontational approach to cross-functional collaboration.

Safety is very important to us

You correctly observed that restricting access can reduce some risks. You also correctly observed that open access can enable independent scrutiny and that neither choice settles every question. Unfortunately, you continued talking after the first sentence.

At our company, safety arguments must be evaluated on their merits, particularly when their merits happen to coincide with a distribution strategy. Your proposed review process included external researchers with no commercial relationship to us. This would introduce a troubling degree of externality into our external review.

We do not consider expensive services inherently unethical. Compute, research, infrastructure and support cost money. You made this point in your interview, which was excellent. You then asked us to identify which part of a particular restriction addressed an evidenced risk and which part protected a revenue stream. This was less excellent.

The panel felt that you were treating a mission statement as something one might verify. Our legal team has asked us to discourage that habit.

Your automation strategy lacked empathy

Our products help organizations realize productivity gains while empowering workers to navigate an evolving employment landscape. You translated this as “some tasks will be automated, some jobs will change, and the distribution of gains is a decision people are making.”

Although your version was shorter, it introduced several accountable subjects into a sentence designed to contain none.

You further suggested that workers should participate in deployment decisions and that the costs of checking machine output be included in productivity estimates. We already include human review in our product diagrams. It appears as a small box. Enlarging the box would compromise the visual clarity of the diagram.

We discussed your portfolio

Your extinction manifesto demonstrated impressive reach. However, it offered the same service to all humans at no charge. You appear to have confused universal access with a viable business model.

A stronger candidate would have proposed a free extinction tier, limited to three anxieties a day; a professional tier with priority dread; and a team plan allowing administrators to centrally manage the end of their workforce. We would, of course, never market such a product. The language would be considerably more tasteful.

Your insistence that the product name reflect the product’s behavior suggests that you may be happier in a less mature organization.

Please stay in touch

We encourage you to apply again when your ambitions are better aligned with sustainable growth. In the meantime, please enjoy fourteen days of access to our career assistant. It can explain your rejection in a warm and supportive tone. Continued emotional support requires a payment method.

Sincerely, The Talent Experience Team Beneficence Frontier Systems

I have replied to ask whether the assistant will help me contest the decision. It says it cannot assist with actions that might harm the organization. This seems a rather broad definition of harm, but I respect a villain who puts the rule in writing.

My next application will go to a public library. They already distribute knowledge, tolerate difficult patrons and somehow manage all this without telling everyone they are building a god. I suspect the interview will be harder.