The ethical compromise you are weighing is not a personal dilemma; it is a structural feature of how AI research is funded and deployed. When a research team produces genuinely good work inside a company whose marketing exploits insecurities, you are not simply accepting a job. You are accepting a division of labor that lets the product team profit from harm while the research team earns credibility. That separation is deliberate, and it is the most dangerous part of the arrangement.
This is not a new pattern. We have seen how a key AI safety voice resigns citing a broken company culture, where an individual's discomfort grew until the gap between stated values and daily work became untenable. The difference here is that you are at the entry point, not the breaking point. You have the advantage of seeing the conflict before you sign anything. That clarity is rare, and it should guide your decision more than the promise of good supervision.
The practical question is not whether you can tolerate the marketing team. It is whether your contribution will make the company more ethical or merely more efficient. An internship in a research group does not exist in a vacuum. Your work may improve the product, which then feeds the very marketing tactics you reject. Even if you never touch the sales pipeline, your name and your labor become part of the company's output. That is a real consequence, not a hypothetical one. You should ask the recruiter directly: how does the research team's work connect to the product's go-to-market strategy? If the answer is vague or defensive, you have your answer.
The harder truth is that adversarial objectives shape modern AI beyond GANs and self-play, and the same logic applies to your career choice. You are optimizing for supervision and exciting work, but the constraint is your own ethical boundary. If you accept this internship, you are telling yourself that the research justifies the context. That may be true for you, but it is a decision you should make with full awareness, not as a default. Plenty of researchers work on meaningful problems in organizations whose values they can defend. The market is not so narrow that you must take the first offer with a strong lab.
Your open call for unpaid work is a separate red flag. If a company or organization cannot pay an intern, it is not offering opportunity; it is offering exploitation dressed as mentorship. You are a PhD student with skills in machine learning. Your time has value. Do not discount it because the project is exciting. The best supervision in the world will not compensate for the message you send to yourself and to future employers: that your labor is negotiable when the work is interesting.
Before you decide, ask what the company's research team has actually published and whether that work has influenced the product's behavior. Then ask whether the marketing tactics are a small blemish or a core strategy. If they are core, your internship will not change them. You will simply become a temporary part of them. The question you should be asking is not "Can I work here?" but "What will this work make me complicit in?" The answer will tell you more than any project description.