By the same token, people could biologically engineer humans to enjoy work, but that wouldn’t make slavery ethical.
As I said, there are things humans can do that robots, machines, and software can’t do. Moreover, they might never be able to do these things without acquiring sentience and hence rights. At that point, you either have to obtain their consent and pay them for their work or resort to slavery. So, there are some jobs that mindless machines may never take over from humans.
Jobs involving research are one example. There are robots that can be used to automate low-level research tasks, but the high-level tasks like thinking about a complex scientific problem or developing a new scientific theory will almost certainly never be taken over by mindless machines (although machines with minds will of course be capable of those tasks).
Just like they do today, companies will need research as one of the raw materials that go into innovation. Today, it isn’t feasible for a large part of the workforce to do research because most of the workforce is needed to run the day-to-day economy. If people stopped showing up to drive buses, sell furniture, and make Mac Pros so they could focus on long-term problems, the economy would collapse. However, if most of the day-to-day economy can be taken care of by robots, that will free people up to work on research. It will no longer be economically unviable.
In the future, most companies might look like Alphabet, which spends 15% of its revenue on research and development and funds forward-looking projects like Calico that will take at least 10 years to come up with its first product. Fundamental and long-term research and development may become one of the biggest sectors in the economy, on par with the service sector today.
This is just one example. If you use your imagination a little, it’s not hard to think of many more. It is far from a foregone conclusion that there will be fewer jobs in the future than today.