The ethical way to proceed is to ask living persons their consent before adding them to a work of genealogy. Just as you wouldn't use a living person's DNA sample without explaining the purpose for which it would be used, and securing their informed consent, you should talk about these issues with a living person whose information you want to include and secure their consent.
The alternative [to dead-naming him] is to not note any [name] change, but I'm afraid that could be confusing to people who don't yet know of the change.
If you are considering 'outing' someone, then you most certainly need to stop and think about what you are doing, and talk about the issues with him. The PFLAG National Glossary of Terms says in their entry Coming Out:
Not everyone is in the same place when it comes to being out, and it is critical to respect where each person is in that process of self-identification. It is up to each person, individually, to decide if and when to come out or disclose.
In the section Disclosing Birth Names, the GLAAD Media Reference Guide - In Focus: Covering the Transgender Community says:
Do not reveal a transgender person's birth name without explicit permission from them.
You talk about not including a "change in gender (from female)". This language usage describes the point of view that someone might take if they were only looking at the historical record, not considering the life of the actual person. Most people wouldn't speak of a "gender change" if an infant was mistakenly recorded as a 1-month old female in a census record but was enumerated as male in all his other census records. While that conflicting evidence needs to be taken into account (resolving conflicting evidence is one of the elements of the Genealogical Proof Standard), most hobbyist genealogists would say that the gender assigned to that member of the household in that first early census must have been a mistake. (As a side note: mis-gendering happens, both when records are created and when they are indexed for computer databases. I've seen genders in the computer indexes that don't match what's on the record recently in the US City Directories and in the census collection on Ancestry, as well as examples like the one mentioned earlier where children were recorded as different genders in multiple census years.)
Record creators in the past may have assumed a gender or assigned a gender that does not reflect the lived experience of that person. Strive for precision and clarity. Avoid usages such as "on the census [the person's name] said". We rarely know who gave the information that appears on a census record.
See the entries in PFLAG's glossary Assigned Gender and Assumed Gender. Consult the living individual about whether or not to include information about the gender assigned at birth, and the language that you could use to describe that information. The trans individual whose personal information is involved can tell you what language usage is acceptable to him. The polite thing to do with any living person is to use the language they prefer.
We don't have to be professional genealogists to benefit from the standards and ethical guidelines used by the pros. The Association of Professional Genealogists Code of Ethics and Professional Practices says, under point 9:
[I agree to] [t]reat information concerning living people with appropriate discretion
The BCG's Genealogists' Code of Ethics says:
I will keep confidential any personal or genealogical information disclosed to me, except to the extent I receive consent to share.
The National Genealogical Society's Guidelines for Sharing information with Others recommends that family historians and genealogists consistently:
• inform people who provide information about their families how it
may be used, observing any conditions they impose and respecting any
reservations they may express regarding the use of particular items;
• require evidence of consent before assuming that living people are
agreeable to further sharing or publication of information about
• convey personal identifying information about living people—such as
age, home address, genetic information, occupation, or activities—only
in ways that those concerned have expressly agreed to;
• are sensitive to the hurt that information discovered or conclusions
reached in the course of genealogical research may bring to other
persons and consider that in deciding whether to share or publish such
information and conclusions.
The Code of Conduct/Ethics of the International Association of Jewish Genealogical Societies says:
Generally, requests made to genealogical researchers should be respected when individuals ask that certain information about
themselves or family members be kept private.
In short, the ethical standards of the day all recommend the same advice that was already stated in the previous answer. Ask living people what information is appropriate for you to include, and ask what language usage is acceptable to them. If your trans interviewee / family member asks you to use the term "gender assigned at birth" rather than talking about a "gender change" then use the requested language.