Except that unmanned missions are not that much cheaper for the same level of capability - robots and probes have limited sensor capabilities, little or no ability to improvise especially not in time critical circumstances, and generally are just physically inferior to humans. For example the opportunity rover averaged 28 feet per day of travel. The lunar rover on Apollo 17 drove further in 5 hours than the curiosity, spirit, perseverance, and sojourner rovers put together. Humans have excellent strength and dexterity, able to easily move or reorient objects weighing many kilograms of arbitrary shape, and overcome arbitrary obstacles. While there are significant costs for this additional capability, the fact is robots simply can't be made to outperform humans at these tasks at the moment. The divide was even worse in the early days of space exploration.

Then, the idea that things would be substantially cheaper without man-rating is also based on shaky assumptions. The fact is large space probes are extremely expensive pieces of hardware and launch windows for certain missions are uncommon, so even if humans aren't on board it is still extremely undesirable to lose a probe to an engineering failure. Indeed while putting a price on human life is extremely taboo, the fact is that in situations where such a price must be determined, it usually comes out to the low tens of millions of USD. Compared to the multibillion dollar costs of many space missions, it makes sense to treat them with the same care as human life.

Finally, the idea that quickly iterating unmanned projects will make human spaceflight safer doesn't make much sense. If you're not using safety systems regularly, and you're not under pressure to improve them, why would one expect to gain useful experience and develop better improved versions? If anything, in the future we should have more capable probes that swing the pendulum further away from human spaceflight.