Eproctoring is a terrible solution to a real problem.
There is cheating at universities. Traditional invigilated exams are fairly effective at detecting certain kinds of cheating, and ensuring a minimum level of competence in some subjects.
I’m aware of a number of different notions of cheating, though I am by no means an expert:
The sort of looking over somebody’s shoulder to see their answers is what comes to mind first, and this is fairly easy to detect through existing means (turnitin etc for written work). This is a common form of plagiarism.
There is collusion, where people try to share answers. There are varying schools of thought on how bad this is. I personally don’t worry too much about it because most assignments are primarily formative rather than summative examinations, and in these cases it is basically peer learning a lot of the time.
The most heinous kind of cheating and the hardest to detect is commissioning. This is basically when students use essay writing services or similar. They will generate an original work of appropriate quality.
It’s almost always possible to detect the kind of blatant cheating people are particularly worried about (commissioning and plagiarism) through sufficient contact time between the students and staff. A brief conversation every now and then with students will promptly reveal blatant cheaters.
There is no magic technological solution to this. Eroding student trust and privacy through invasive surveillance will inevitably lead to worse outcomes.
The solutions to this are to set better assignments, which require thoughtful answers, rather than scantron-like quizzes where answer keys can be shared, high levels of contact time to get to know students, and occasionally controlled exams to ensure nothing is slipping through the cracks. There’s a big difference between teaching and certification, and it’s very hard to balance those needs.
Anything summative needs to be marked anonymously (or at least pseudonymously), and often needs to be done under controlled conditions.
It is not possible to detect a sufficiently advanced cheater (particularly people commissioning work) working in an environment they control, on a device they control. The cheaters will beat you. They do in video games, and they will in academic pursuits too.
The people selling this are preying on a legitimate fear of cheating that universities have, but their solutions are snake oil, and degrading to the majority of students who are honest and want to learn. Better teaching solves many of these problems, and individual projects with vivas are an extremely effective solution for where you need higher assurance. We can go back to invigilated exams if we desperately need to after covid.