Goldman Sachs has issued a strict warning to young applicants: avoid using artificial intelligence tools during the hiring process. Instead, the $176 billion financial giant advises candidates to thoroughly research the company before attending interviews.

Goldman is not alone. Other major companies, including Anthropic and Amazon, have also made it clear that the use of AI during interviews may result in immediate disqualification.

This policy was disclosed by Goldman Sachs' recruitment team for the Private Investing Academy program in the EMEA region. In an email to students—reported by eFinancialCareers—the bank emphasized that candidates preparing for interviews via HireVue, the AI-powered assessment platform used to evaluate behavioral skills, should focus on studying the company’s financial performance, core values, and guiding principles. Relying on artificial intelligence is explicitly discouraged.

“The use of external sources, such as ChatGPT or Google, is strictly prohibited during the interview,” the email stated.

Goldman interviews typically allow only two minutes for a response with just 30 seconds of prep time. This tight timeframe makes it nearly impossible to generate a meaningful AI response during the interview. More importantly, AI-generated answers often lack authenticity and personalization—two factors that can seriously undermine a candidate's chances.

There’s a notable irony in these policies: while Goldman bars candidates from using AI, it provides such technologies to nearly 50% of its workforce. Still, this isn’t unique to Goldman. Many companies are navigating inconsistent AI policies as they shape their broader AI strategies.

Tech companies are also taking a hard line.

Anthropic, valued at $61.5 billion, recently emphasized in its hiring campaign: “Do not use AI assistants. We want to see your genuine interest and real communication skills.”

Amazon, valued at $2 trillion, issued an internal directive to its recruiters earlier this year: “Candidates found using AI should be removed from the process. AI tools mask true capabilities.”