My nephew started infant self-rescue classes when he was 6 months old. We filmed his progress, and the results show that babies are capable of much more than we might think.
We don't typically demand much of infants. Making them comfortable and meeting their needs has always been the obvious priority. But in some cases, it's worthwhile to make babies uncomfortable to reduce their risk of future harm. Vaccines are one example. For my family, self-rescue classes are another.
My nephew was 6 months old when he began training with Infant Swimming Resource, a network of certified instructors who teach babies aquatic survival skills. These lessons are the probably most challenging thing he'll have to do for years to come. They can also be hard to watch.

By the end of week 1, he seemed to understand that his body should face upwards, but struggled to properly orient his hips and neck to stay afloat.
The classes work from the assumption that you can shape an infant's behavior with repetition and reinforcement. They do not use any floating devices, which tend to orient children vertically in the water — a position that would spell doom for a baby who slipped into a pool.
The primary skill taught is finding a float and maintaining it until help arrives. Unfortunately, to learn the proper body position, he had to first experience what doesn't work — what gets water in his nose and what sends him back under the surface. It was not fun. But after three weeks of daily classes, my nephew had been successfully trained to float.
By the end of week 3, he wasn't happy about it, but he could find his float and stay there for a while. He was tested in clothing before he graduated.
The classes are Monday through Friday for 10 minutes per day, and they cost $85 per week. They are not recommended by the American Academy of Pediatrics, the organization of pediatricians that publishes evidence-based advice. There have been anecdotal accounts of ISR classes saving lives, but because nobody has yet published a rigorous evaluation, and because ISR training could conceivably make parents less careful with supervision, AAP does not endorse them.
It's worth noting that until 2010, AAP said that children were not developmentally ready for swimming lessons until age 4. But in 2009 a case-control study suggested that swimming lessons for children ages 1 through 4 could significantly reduce drowning risk. Still, as AAP emphasizes, pool fences and adult supervision should remain top priority.
Having a baby in a pool is also not without health risks. Water can transmit bacteria that cause diarrhea, and chlorine has been linked to respiratory problems, particularly in families predisposed to allergies.
Whether these classes make sense for your baby may depend on how frequently they access pools or bodies of water. The national drowning death rate for children ages 1 through 4 was around 3 per 100,000 annually, between 1999 and 2010. But the drowning rate is higher in Arizona, where my nephew lives, and in many of the Southern states. My sister's home has a swimming pool, as does her mother-in-law's (along with 43 percent of homes in the Phoenix area). That's why my family decided not to wait for more data.
He's now 1, and he's learning how to swim underwater. He returns to his float when he needs air and rest.
Now that my nephew has passed his first birthday, he's learning how to swim underwater. It will still be a couple of years before he's able to lift his head above the surface while doggy paddling, so instead they're teaching him to swim toward the stairs, roll into a float to rest and breathe, and then flip back over to keep swimming. None of this makes him "drown-proof," but it could buy him more time in an emergency situation and give him a fighting chance to save his own life.