What are the ethical considerations when publishing a cheating technique?

I can’t really tell you what to do with that vague level of detail, but I can give you an example.

There seems to be a team at the Ben-Gurion University in Israel that regularly comes up with obscure ways to transmit data to devices not connected via any network connection.

Stuff like:

• DiskFiltration: Data Exfiltration from Speakerless Air-Gapped Computers via Covert Hard Drive Noise,

• SPEAKE(a)R: Turn Speakers to Microphones for Fun and Profit,

• and many many more.

The only uses I see for that stuff is to either help malware authors and intelligence agencies or make people aware of this communication channels to increase security.

If your findings are anything like this, I don’t see a problem.

Update After you clarified it's about a method to remotely stream answers of a test directly into the retina without anyone else noticing:

There are a lot of uses for such a technique which are not related to cheating. Pick some of them as motivation, and then later in your paper describe how this could also be used for cheating. Maybe give hints what some mitigation measures could be.

I'd go by precedence set by large companies such as Google. If you know of specific companies/organisations who would be affected by it, give them a notice period before you publish. After that, publish as you normally would.

Problems don't get resolved unless people are aware of them. If your cheat comes from the actual process, then the process needs to be changed. If it's not, then you've invented a valuable new technique.

Are there any less ethically dubious applications of your algorithm? I take it this is an incidental finding if you find cheating so unethical. If your main intention is to publish the method, perhaps it world be better to emphasise these other applications?

Either way, I'd recommend publishing anyway. Aside from prescedence and contributing to the literature being good for your career, it may also benefit those working to detect and reduce such unethical conduct. As you say, it's very possible someone less hesitant to use it will discover it so people who could prevent it being misused may as well be aware of it too.