Recently The Daily Mail reported on a
study that revealed AI systems
are likely to impact around 80 percent of US jobs.
The research conducted by Artificial Intelligence
(AI) developers - OpenAI and the
University of Pennsylvania also discovered that 15 percent of all worker tasks
could be accomplished more efficiently by AI and with identical quality.
Concerns
that AI apps could replace human workers caused headlines last November after
the launch of the advanced AI App ChatGPT.
'GPT' in ChatGPT stands for generative pre-trained transformer, which
enables applications to learn from past records and produce human-like text.
ChatGPT can perform tasks such as writing emails, columns, resumes, etc.
It
doesn’t just write, it can even mimic the writing styles of other authors.
Perhaps
you can summon the ChatGPT to write columns in the style of Ernest Hemmingway.
The app purportedly mimics Hemmingway in its construction of sentences, choice
of words, and ideas without plagiarizing directly from any of Hemmingway’s
works.
Back
to the University of Pennsylvania
study.
1,016
occupations and their 19,265 tasks were analyzed to evaluate which tasks could be
done by AI applications.
The
overall findings show that at least 10 percent of work tasks are affected
by the introduction of the Large Language Model (LLM), while approximately 19
percent of workers may see at least 50 percent of their tasks impacted'
LLM in AI refers
to AI models that can generate language texts on any topic or domain by studying large volumes of data.
The
study also evaluated qualities such as critical thinking, listening, and
speaking, and compared how the AI ranked for those qualities for specific jobs.
The inference was that 86 occupations were "fully exposed" to AI i.e. the tasks can be achieved by the AI apps entirely.
The
most at-risk are white-collar jobs, such as mathematicians, accountants,
writers, and occupations.
However, jobs that need tactical or strategic thinking or innovation such as marketing strategists, etc. scored poorly with just 15 percent of tasks being replaced by AI.
Blue-collar
jobs such as dishwashers, electricians, barbers, etc. are obviously safe from the
AI application, perhaps until robot technology becomes more affordable.
AI is already being used.
A few days back an Indian Court used ChatGPT as one of the inputs while ruling on an order for bail in a murder case.
So can
humans be replaced by AI
Let's
revisit the Hemmingway example.
The
AI can mimic Hemmingway's style of writing and perhaps his ideas by studying Hemmingway's previous works.
But
AI obviously cannot predict human beings changing their mind following personal
experiences or global events.
Perhaps
Hemmingway’s style of writing and opinions would have evolved had he lived
longer, but the app cannot fathom that eventuality.
The AI may be able to perform all tasks performed by an accountant more efficiently than a human being.
But the human accountant could devise easier practices of bookkeeping and apply his acumen to ensure that his client pays the least amount without violating tax
Where
the AI app could fail is innovation.
But for repetitive tasks, AI apps could be a blessing
In such circumstances, many jobs will be at peril.
What about
AI bot bias and control?
We
learned that lesson from big tech's usage of social media.
The
likes of Twitter and Facebook wriggled into personal and professional lives and
even became the official channel of communication for governments.
Once
they developed a monopoly, they seized control. Rather than reflecting
public opinion, they began manipulating the discourse to benefit certain vested interests.
A
recent report revealed
that the ChatGPT struggled to define a woman, praising Democrats but not
Republicans, refused to defend fossil fuels, and claimed nukes are less
dangerous than racism.
If
humans become overly reliant on AI apps, there will come a time when they take
over without humans noticing.
The
impact could be information suppression - the AI app doesn’t present information that
expresses skepticism about man-made climate change, the COVID-19 vaccine, etc. This is already happening with Google.
The
app could be used by governments to control and monitor citizens.
If the app controls your home security system, the government could collude with the developers or administrator to alter the app to keep you locked indoors because you are unvaccinated, the app merely has to look through the government vaccination database to find that information. Perhaps you are locked at home on election day because the app knows you will vote for Trump based on your social media behavior.
Perhaps
the application monitors your conversations and notifies government officials
of ‘extremist’ ideas, Amazon’s Alexa could be
tweaked a bit to have that capability.
We are already aware of deep fake technology where AI manipulates videos and can even replace the likeness of one person with another.
The
following video shows how deeply fake technology replaced the images of the actor who played Han Solo in Solo:
A Star Wars Story (2018) with that of Harrison Ford.
https://www.youtube.com/watch?v=5WE3ztdy7HQ
With
some refinement, this technology could be used to alter video footage to incriminate an
individual in a court of law.
Perhaps the technology can be used in a political campaign to make it appear that a political opponent has used offensive invectives or has insulted his supporters.
These
AI apps also present a serious risk to educational institutions.
New York University explicitly prohibits the use of ChatGPT for completing assignments and quizzes. Students were warned by professors on their first day to not use the AI tool to cheat, according to Vice.
Big
tech also sees AI apps as a threat to their monopoly. Gmail co-creator Paul
Buchheit predicted that ChatGPT would eliminate the need
for Google's search engine altogether, thus disrupting the
company's main source of revenue.
There
are some who think that governments must intervene to prevent this AI takeover of businesses that will destroy livelihoods.
What we need to be vigilant about is governments using AT technology to bolster their power.
But it is impossible to suppress technical advances in the marketplace.
Once
upon a time, something similar was said about computers yet today the personal
computer is an essential component.
Companies,
especially smaller businesses can save valuable funds with the judicious
use of AI apps.
We
have to remember that since the invention of the wheel, humans have been told
that their livelihood and existential purpose are under threat.
Yet
our species has managed to evolve and keep up with the changes.
Doubtlessly
there will be challenges even with AI and humans will have to learn to adapt.
AI is like any other tool, it is up to the human to decide the level of control and the manner of usage.
Also appears on AT
Comments
Post a Comment