We don’t know.
The difficulty is that there are no fossils of individual bacteria and viruses and the kinds of thing we presume were the first living things (perhaps an RNA molecule in a lipid bubble). Their chemical components would long ago have degraded - and even the rocks that are old enough to contain any fossil evidence have almost entirely been subducted and melted back into the Earth’s core.
Here is a photo of the very earliest evidence we have for life:
This is a bunch of tiny filaments, blobs and tubes in Canadian rocks dated to be up to 4.28 billion years old. They have the right chemical composition to represent a once-living thing - and the structure seems right.
But this wasn’t the VERY first living thing…it’s just the oldest one we have evidence for.
The FIRST living thing would have been a simple self-replicating molecule. Something which, by virtue of very mundane chemical processes makes copies of itself from ingredients just floating around in the ocean (or wherever it originated).
The creation of that first ever self-replicating molecule is called the "abiogenesis event"...and science doesn't know much about how it happened.
It’s essentially impossible to ever find conclusive and direct evidence of the very first living thing, so we’re left with more of an experimental approach:
This is not a comfortable way to proceed - because all we’d ever be able to show is that spontaneous generation of life is possible via our best-guess route. We still couldn’t tell whether there might be other possible routes - and which of those routes actually happened.
Another problem is that of probability.
The approach above works well if whatever happens in step (4) is a high-probability thing…we set up the conditions in the lab - wait a month, then look for life and find it.
But if the probability of this spontaneous generation of life is very low - then we’ll never see it in a laboratory experiment - even if it runs over a square mile of early-earth environment and we let it run for a century.
Here on Earth, there were perhaps a hundred million square miles of ocean - and it could have taken 200 million years of random recombination of amino acids for just one suitable event to occur to create that first self-replicating molecule.
It could be even less likely than that. There are probably billions of suitable planets in the Milky Way and at least 100 billion similar galaxies in the observable universe. Life only had to appear once in all of those places (and that’s where we live). If the universe is truly infinite (which could easily be the case) - then no matter how statistically unlikely it is for life to arise spontaneously - then it would definitely happen…and again, that place was here.
If it’s that unlikely - then we may never see this happen in a laboratory.
WHAT WE KNOW NOW:
In the “extremely improbable” case, we’ll never be able to show direct evidence for the creation of life.
We’d be left with exactly what we know now:
If we calculate the probability of that spontaneous creation of that relatively complex molecule - even with the most pessimistic assumptions…we know that the probability cannot be zero. If it isn’t zero then somewhere in all the oceans of all the planets of all of the stars in all of the galaxies and over a span of 14 billion years…life was inevitable - and we are the result.
So in a sense, we already know the entire story.
The trouble is, for the general public’s understanding, it’s an uncomfortable answer. It would be nice to show that the probability of this happening is high enough that life happens easily...but if it's not, then it's a harder argument still.
We’d absolutely want to be able to create life ‘spontaneously’ in the lab…but there are no guarantees that this is plausible.