Blog

Your PhD Student Can't Do Research and You're Not Their Supervisor: What Now?

When the person doing the supervising has no official role and the student is not progressing, walking away is easy. Knowing what to tell the real supervisor is not.

Binary Buddies · 5 min read · 0 views
A research office after hoursAn empty office at dusk: a whiteboard of half-finished diagrams with one box circled, a glowing laptop on a desk, two chairs pulled apart, and rooftops outside the window.

A postdoc in a research group in Europe put a question to Academia Stack Exchange that plenty of lab members will recognise from the inside. For more than a year they have been informally supervising a PhD student who, in their judgement, cannot do research. Faced with any problem, her approach is to ask AI and proceed from whatever comes back, programming included. In the weekly update meeting she cannot say why she took an approach, what the next steps are, or why they would matter.

The awkward part is the org chart. The official supervisor handed the postdoc essentially all responsibility for the student, is kept regularly updated, and will likely agree to whatever the postdoc suggests. Several conversations over several months produced some improvement, but not enough, in the postdoc's view, to get her a PhD. In their country it is not easy to dismiss a doctoral student for these reasons, nobody else in the group wants to supervise her, and the postdoc no longer wants to.

Walking away is technically easy, because the role was never formal. The hard part is what to say when the supervisor asks what happens next. The two most-voted answers on the thread take different routes, and both have more to say about the adults in the room than about the student.

Key takeaways

  • Being “regularly updated” is not the same as understanding: diplomatic phrases like “some concerns” can make a failing student sound merely below average.
  • The decision belongs to the official supervisor; the postdoc's job is to make sure it is taken on an accurate picture.
  • If the postdoc is asked to carry on, the top answer calls for written, objective pass/fail criteria, an agreed exit, and a student who knows both.
  • A blunt conversation plus a one-week deadline for a plan turned one struggling master's student around, but only with honest self-reflection from the supervisor.

The complaint isn't the AI, it's the missing why

Look closely at what the postdoc is actually measuring. The complaint is not tool use as such. It is that after a year the student has not acquired the skill of independent problem solving, and the AI-first habit is where that shows. The test is applied every week, and it has three parts: why this approach, what comes next, and why would that be relevant. Those three questions come close to a working definition of directing your own research, and she repeatedly cannot answer them.

The postdoc also says there are further reasons she lacks the attitude of a problem solver needed in their field, which they decline to list and ask readers to take on trust. That matters for anyone weighing the case from outside: the visible evidence is the meeting pattern, and the rest is one person's assessment. It also matters for whatever comes next, because any formal step will need something more concrete to point at than an attitude.

1+ yearStudent's time employed in the group
28Votes on the top answer
1 weekDeadline set for a turnaround plan

Unofficial supervision, official consequences

The arrangement in this group splits authority from work. The official supervisor holds the title. The postdoc holds the weekly meetings, the judgement calls and, in practice, the student's day-to-day progress, while the supervisor is updated and is likely to sign off on whatever the postdoc proposes.

That gives the postdoc a great deal of influence and no formal standing. They can stop giving their time whenever they choose, but stopping is not a plan: the student remains employed, no one else is willing to take her on, and the supervisor will still want an answer. The postdoc's own goal is narrower and sensible: they have a really good relationship with the supervisor and want to walk into that conversation prepared.

Who supervises whomThe official supervisor delegates responsibility to a postdoc, who holds weekly meetings with the PhD student and reports back; no one else has volunteered to supervise.Official supervisorPostdoc, unofficialPhD studentDelegates responsibilityRegular updatesWeekly meetingsNo other volunteer
All the day-to-day supervision runs through the postdoc, and there is no one behind them to hand over to.
The weekly meeting roomA small meeting room with a round table, two chairs facing each other, sunlight falling through a window, and a wall calendar with the same weekday marked in every row.

First, check that the supervisor heard “failing”

The highest-voted answer, with 28 votes, opens with a question the postdoc may not have asked: the supervisor has been regularly updated, but have they really internalised what they were told? People in the postdoc's position tend to be diplomatic, reaching for euphemisms like “performance issues” and “some concerns”. The postdoc leaves believing they have made clear the student is failing; the supervisor leaves believing she is merely below average.

So the first move is not a proposal but a correction: make sure the supervisor knows the student is failing and has been for some time. Only then does the answer hand the decision back to the person who owns it. The supervisor may agree that the student has already failed, may make other arrangements for her, or may ask the postdoc to continue working with her.

The euphemism gapDiplomatic phrases from the postdoc are heard by the supervisor as merely below average, while the reality is that the student is already failing; the fix is to say failing plainly.Same update, two readingsThe postdoc saysThe supervisor hearsThe reality“Performance issues”“Some concerns”Merely below averageAlready failingFix: say “failing” plainly
Polite wording lands in the middle box; what the supervisor needs to hear is in the coral one.

If asked to continue, put the exit in writing

The third outcome is the one the postdoc least wants, and the same answer treats it as a negotiation rather than a favour. It recommends pushing for a clear, written agreement with both the supervisor and the student, setting objective success and failure criteria. The supervisor should also agree up front that the postdoc will stop supervising if the criteria are not met, and the student should be told that as well.

The value is that it replaces an argument about attitude with a check against terms everyone saw in advance. It also protects the student: she learns exactly what is expected and what follows if it is not delivered, instead of discovering a verdict reached in conversations she was never part of.

The supervisor's three optionsOnce the supervisor understands the student is failing, they can declare failure, make other arrangements, or ask the postdoc to continue; continuing leads to written pass/fail criteria, known to the student, with the postdoc stepping away if they are missed.Supervisor hears “failing”Declares failureOther arrangementsAsks you to continueWritten pass/fail criteriaCriteria missed: you stopStudent knows the terms
Only one branch keeps the postdoc involved, and that branch has its end point agreed before it starts.

  1. Use the word “failing”

    Drop “concerns” and “performance issues”. Say the student is failing and has been for some time.

  2. Bring the record

    The repeated conversations over several months, the improvement that did happen, and why it still falls short of a PhD.

  3. Let the supervisor decide

    Declaring failure, making other arrangements or asking you to continue are their calls, not yours.

  4. If you continue, write the criteria

    Objective success and failure criteria, agreed in writing with both the supervisor and the student.

  5. Agree the exit in advance

    The supervisor confirms you stop if the criteria are missed, and the student is told so.

A plan at dawnEarly sunlight through a window onto a desk; on the wall a pinned checklist with some boxes ticked and a strip of seven days, five of them filled.

The blunt route, and its price

The second answer, with 20 votes, comes from someone who supervised a master's student with similar-sounding trouble, and it takes a more direct line. They laid out all their problems with the student's work and attitude, without being impolite but also “without gloves on”. That was new to the student: until then, people had always tried not to hurt their feelings and motivation. The answerer is candid that it was not a nice experience for either of them.

Structure followed. The student got one week to come up with a plan for the next few weeks, with suggestions such as SMART goals, general directions, the supervisor's expectations and topics to review, and a clear statement that supervision would end if the plan was not taken seriously. During that week the supervisor stayed available for specific questions. It worked: the student stuck to the plan, then developed it further on their own, and graduated.

The one-week turnaroundA blunt talk leads to one week to write a plan, with questions still answered and supervision ending if the plan is ignored; the plan is followed and the student graduates.One master's student, one weekBlunt talkOne week to planPlan followedGraduatedQuestions still answeredElse: supervision ends
The deadline is the centre of the method: support stays on during the week, and the consequence is stated before it starts.

Escalate to the supervisor

The 28-vote answer, written for this exact setup.

  • Restate the situation as failing, not concerning
  • Let the official supervisor choose the path
  • If continuing: written, objective criteria and an agreed exit

Blunt talk and a one-week plan

The 20-vote answer, from supervising a master's student.

  • Lay out every problem, politely but plainly
  • One week to write a multi-week plan, with help on request
  • Supervision ends if the plan is not taken seriously

The same answer attaches a condition that is easy to skip. The method demands a lot of reflection from the supervisor: about their expectations, their abilities as a mentor, what in the situation is on them, and the environment in which it is all playing out. It is not, the answerer stresses, simply ordering a student to produce a plan or be fired.

What neither answer settles

The success story involves a master's student, not a doctoral one, and a supervisor with formal authority, not a postdoc whose authority is borrowed. “I will no longer supervise you” carries real weight from the person whose name is on the supervision. From an unofficial supervisor it is true but incomplete, because the student would still have an official supervisor and an employment contract.

The employment question is the largest gap. The postdoc says it is not easy in their country to dismiss a PhD student for these reasons, and the thread does not set out what process would apply instead. Anyone facing this in Europe should read their university's doctoral regulations and progress-review rules, and talk to the graduate school or HR, before treating stepping back as the end of the story. Whether the AI-first habit is the cause of the problem or only its most visible symptom is also left open.

Can an unofficial supervisor simply stop supervising a PhD student?

In this case, in practice, yes: the postdoc holds no formal role and can stop giving time whenever they choose. But the student, her contract and the official supervisor all remain, so stopping without a plan just moves the question up a level.

Should the student be told she is failing?

Both answers reject vagueness. The top answer has the student know the written criteria and the consequence of missing them; the other describes telling the student every problem directly, after others had shielded them.

Does relying on AI mean a PhD student can't do research?

Not by itself. The problem here is that after more than a year the student cannot explain her approach, her next steps or why they matter; asking AI first is how that shows up.

Can a PhD student in Europe be dismissed for poor progress?

It depends on the country and institution. In this postdoc's country it is not easy for these reasons, so check the local doctoral regulations before planning around it.

This article draws on PhD student is unfit for research, what do to? and answers to it (1, 2), on the Stack Exchange network, licensed under CC BY-SA 4.0.

  • #academic supervision
  • #europe
  • #failing phd student
  • #phd
  • #research independence
  • #supervision

Comments

Add a comment
Sign in with Google to join the conversation

Keep reading

Education · 5 min

Determinant of e^A e^B e^-A e^-B: Trivial for Matrices, Substantive for Operators

Education · 5 min

A Closed Form for ∫₀^∞ (x coth x − 1) Li₂(e^(−x)) / x dx, Derived Two Ways

Education · 5 min

Closed form of ∫₀^∞ x^(2n) arctan(x)/(e^(aπx)+1) dx via negative-order polygamma