Did you know AI Might Surpass Humans in All Tasks by 2047
The rise of AI has been so unbelievably rapid that within the span of
just one year it managed to become perhaps the central topic of
discussion in the world of tech as well as in day to day life. The main
question that people tend to ask here is how effective AI will be at
replacing humans in a wide variety of tasks. It turns out that
researchers at AI Impacts tried to figure out the likelihood of this
occurring in the future.
The researchers at this organization
collaborated with the University of Oxford and the University of Bonn to
survey 2,778 experts who have published their thoughts on AI. According
to these authors, there’s a 10% chance that AI will outperform humans
in every single task you can think of over the course of the next three years.
If
the current trend persists, these experts say that the chances of it
surpassing human beings by the year 2047 is around 50% with all things
having been considered and taken into account. As for human occupations
in particular, researchers concluded that there’s a 10% chance of all of
them becoming automated by the year 2037.
Translating text, recognizing objects after seeing them a single time,
writing basic code in Python, writing fiction that could reach the New
York Times bestseller list, creating a payment processing site, and even
creating large language models of their very own could become a
reality. AI is advancing far more quickly than anyone could have
expected, which makes it rather necessary to figure out where things
might go from here.
It bears mentioning that these are perhaps
the most pessimistic predictions that anyone could make. Optimistic
predictions have also come about, with many saying that AI will
complement jobs rather than replace them entirely, but the fact of the
matter is that the pessimistic outlook is no less likely than the
optimistic one. Indeed, 81% of survey respondents believe that AI will
be able to talk like human experts within the next 20 years, which would
make the respondents themselves obsolete all in all.
m