By Brian O'Donovan
Work & Technology Correspondent
ChatGPT is one year old today.
The powerful artificial intelligence (AI) tool that has taken the internet
by storm was launched on 30 November 2022.
The online chatbot can provide fast, human-like responses to questions and
can also write essays, scripts, poetry and even jokes.
It is a "large language model" trained to produce text using vast amounts of
data from the internet.
In the weeks after its launch, ChatGPT became the fastest-growing consumer
application in history, according to a UBS study.
Data from analytics firm Similarweb shows that on the day it was released,
ChatGPT attracted 153,000 visits.
By the end of the first week it was at 15.5 million visits, growing to 58
million by the second week.
It was estimated to have reached 100 million monthly active users by
January, just two months after its launch. In May this year, it hit 1.8
billion monthly visits.
It meant that the company behind ChatGPT, OpenAI, saw its website become one
of the most visited domains in the world.
Rivals quickly followed
The launch of ChatGPT represented a massive leap in artificial intelligence
and companies quickly moved to incorporate the technology into their systems
and to develop their own versions.
In February, Google unveiled its rival chatbot Bard.
Microsoft invested billions of dollars in OpenAI and incorporated ChatGPT
into its Bing search engine.
In July, Facebook and Instagram parent company Meta, announced the launch of
Llama 2, its open source large language model.
Earlier this month, Elon Musk unveiled "Grok", a ChatGPT rival with a
"rebellious streak" according to the tech billionaire.
When it launched last year, ChatGPT would only accept text questions and its
database contained information up to September 2021.
There have been updates in its first year of operation. ChatGPT is now
trained up to April 2023 and has access to Microsoft's Bing search engine
for more recent information.
It can also accept image-based requests.
But despite these improvements, ChatGPT is not perfect, something which its
creators admit to.
"ChatGPT can be a helpful tool, but it's not perfect," according to an
article on the OpenAI website.
"Sometimes, ChatGPT sounds convincing, but it might give you incorrect or
misleading information, often called a hallucination," the article added.
"It can even make up things like quotes or citations, so don't use it as
your only source for research."
OpenAI warns that its chatbot can misrepresent different sides of an
argument and does not know everything.
It is also not free from biases and stereotypes.
"Bias mitigation is an ongoing area of research for us, and we welcome
feedback on how to improve," according to OpenAI.
"The model is skewed towards Western views and performs best in English."
The article warns that ChatGPT can reinforce a user's biases over the course
of interaction.
"For example, the model may agree with a user's strong opinion on a
political issue, reinforcing their belief," OpenAI said.
ChatGPT or CheatGPT?
In the weeks following the launch of ChatGPT, there were fears that it would
become "CheatGPT" and would be used by students to write their essays and
assignments for them.
Dr Roisin Lyons, Associate Professor of Entrepreneurship and Innovation at
the Kemmy Business School in the University of Limerick, has researched the
impact of ChatGPT on education.
She said that in the last year, since the launch of the technology, many
educators have had to change the way they teach.
"A lot of teachers that had tended to focus on the output itself, the final
project, have started to look at the draft stage and have started to engage
with the process of how students get from A to B, which has been really
insightful for them," Dr Lyons said.
"Other teachers and educators have moved to more face-to-face assessment,
whether that be in-person tests or interviews, really trying to connect with
the students and their learning," she added.
"There have been positives and negatives for education but I liken it to
when the calculator came along, we all had to navigate through that," she
added.
But what about those dire warnings that there would be widespread cheating
across academia after ChatGPT came on the scene?
"Personally, I teach innovation and I've had really interesting debates with
students about what they consider cheating and how they embrace these
tools," Dr Lyons said.
"You self declare what you've used and we can investigate what has been done
through a number of different tools. So they are using it, but it isn't as
dire as the cheating situation as we had once thought," she added.
Dr Lyons said that fewer students are using it for final assessments and
instead turn to ChatGPT for help structuring projects and articulating
points or to give a first level idea of how to approach a topic or a
subject.
"Some students don't like it all and have found that there's too much heavy
lifting and prefer to do things themselves," she added.
Dr Lyons said that using ChatGPT has benefitted some students'
communications skills.
"We have found that students that are using it tend to be very succinct,
they have a good command of their English and they have a good command of
how to give instructions," she stated.
"This is a skill and a new area of communication that will help them as
managers and delegators in the future when it comes to articulating
themselves coherently," Dr Lyons said.
Embraced by Irish workers
One year after its launch, it is clear that ChatGPT has been embraced by
Irish workers.
A recent study by Deloitte showed that 300,000 people in Ireland have used
AI at work.
Among those who were aware of the technology, over three in five believed
that Generative AI will reduce the number of jobs available in the future
and almost half of respondents said they are concerned that Generative AI
will replace some of their role in the workforce in the future.
Deloitte's research also found that of those who have used Generative AI,
more than one in three believe it always produces factually accurate
responses, and 31% agree that its responses are unbiased.
"By far the most popular purpose of using Generative AI is for personal
purposes while 34% of respondents use it for education," said Colm
McDonnell, Partner Risk Advisory with Deloitte.
"It is clear from these responses that the use of Generative AI will only
increase with time and greater adoption," Mr McDonnell said.
Resignations, promotions and pay rises
A recent study from Workhuman showed that more than a quarter of Irish
workers have used AI tools to help them ask for a promotion and 20% have
used it to ask for a raise.
The Human Workplace Index also showed that almost a quarter of employees
have used AI to help them resign.
Some 40% of Irish workers have claimed work carried out by AI as their own
but of them, half were caught out by a colleague or manager.
70% of respondents said they believe AI will lead to better job satisfaction
and a better work-life balance, while the same proportion believe it will
create new job opportunities. This is higher than the 42% who believe AI
will eventually replace their job.
Trouble at OpenAI
Earlier this month, the board of OpenAI fired CEO and co-founder Sam Altman
in a move that sent shockwaves through the tech world.
The company gave few details as to why he was sacked but said it had "lost
confidence" in his leadership, accusing him of not being "consistently
candid in his communications".
The board also said it needed to defend the companys mission to develop AI
to benefit humanity.
There was an angry reaction from staff, who threatened mass resignations
unless the board stepped down and reinstated Mr Altman.
After he was fired he was quickly offered a job by OpenAIs financial backer
Microsoft who wanted him to head a new research team.
Days later, in another twist, it was announced that Sam Altman would be
returning to the role of CEO and that there would be a partial restructuring
of the board that dismissed him.
It was major corporate drama from the firm behind dramatic leaps in the
advancement of AI.
As the first year of ChatGPT draws to a close, serious questions have been
raised about the company that delivered the technology to the world,
questions that even its all-knowing chatbot would struggle to answer.
RTÉ News.
