Technology CEOs Worry That AI Offers A “Risk of Extinction”

Leading computer scientists and technologists have cautioned that the “risk of extinction” posed by artificial intelligence necessitates international action.

The Centre for AI Safety, a research and advocacy organisation based in San Francisco, released a brief statement on Tuesday that quoted a number of AI experts and other prominent figures.

The signatories include prominent individuals such as neurologist Sam Harris and musician Grimes, as well as industry leaders such as Audrey Tang, Taiwan’s digital minister, and Geoffrey Hinton, dubbed the “godfather of AI” and chief executive of OpenAI.

AI-Assisted Chatbot

The National Eating Disorders Association, based in the US, is in the news after laying off all of its employees and substituting Tessa, an AI-assisted chatbot, for them.

This took place only four days after the six paid workers, who were in charge of 200 volunteers, successfully organised a union. Coincidence? Absolutely, Neda replied; it was a long-awaited reform unrelated to unionisation.

In a recent interview, CEO Arvind Krishna stated that the technology giant would stop hiring for “back-office jobs” in the future years and automate those roles.

All of this may sound incredibly gloomy, yet there are many AI believers willing to promise us that artificial intelligence will actually improve things.

Humans will have more free time to laze around and write poetry in the sun because technology will take care of all the tedious chores.

Greater Emphasis On AI

The alert comes in response to an open letter signed by Elon Musk and other well-known individuals in March that demanded a six-month moratorium on the creation of AI more sophisticated than OpenAI’s GPT-4.

Rapid AI development has led to worries about potential negative effects on society, including widespread job losses, copyright infringement, the spread of false information, and political unrest.

Since then, the number of cautions regarding the possible risks of AI has risen.

Promoting The Dangers Of AI

Geoffrey Hinton, a well-known computer scientist, left his position at Google last month so that he could devote more time to promoting the dangers of AI.

Altman urged lawmakers to swiftly create rules for AI technology at a testimony before the US Congress earlier this month, and he suggested a licensing-based strategy.

The US and other nations are rushing to write legislation that strikes a balance between the need for regulation and cutting-edge technology.

By the end of the year, the European Union intends to approve legislation that would divide artificial intelligence into four risk-based groups.

Chinese lawmakers passed rules overseeing deep fakes and demanded that businesses register their algorithms with regulators as part of their efforts to control AI.

Beijing has also put up stringent regulations to limit the publication of politically sensitive content and demand developer consent before disseminating generative AI-based technology.

What is ChatGPT from OpenAI? We Break it Down Here

A chatbot with artificial intelligence was created by OpenAI and launched in November 2022 under the name ChatGPT. ChatGPT made its debut as a prototype and quickly gained popularity for its thorough responses and clear responses in a variety of subject areas.

Computer experts like Geoffrey Hinton and Yoshua Bengio have expressed concerns about ChatGPT and GPT-4, including the possibility that future AI systems would outperform humans in intellect, pursue the wrong objectives, and pose existential threats.

Reinforcement learning from human feedback, a method for fine-tuning, combined supervised learning and reinforcement learning. Both strategies make use of human instructors to help the model perform better. Users of ChatGPT provide data that OpenAI uses to train and improve the service. Users can give ChatGPT responses an upvote or downvote and leave extra feedback in a text field.

Human Conversationalists

ChatGPT is flexible even though its primary job is to resemble human conversationalists. It has the ability to create and debug computer programs, compose business proposals in the manner of well-known CEOs. write songs, teleplays, fairy stories, and student essays; respond to test questions; make song and poetry lyrics, translate and paraphrase texts, imitate an ATM, play games like tic-tac-toe, and emulate a Linux operating system.

Compared to its predecessor, InstructGPT, ChatGPT makes an effort to lessen negative and dishonest comments. In one instance, ChatGPT acknowledges the counterfactual nature of the question and frames its response as a hypothetical consideration of what might occur if Christopher Columbus came to the U.S. in 2015. It does this by using information about Columbus’ voyages and facts about the modern world, including modern perceptions of Columbus’ action.

OpenAI declared in March 2023 that it would accept ChatGPT plugins.

Regardless of actual comprehension or factual substance, human reviewers favoured longer answers while training ChatGPT. In one instance, ChatGPT produced a rap suggesting that male and white scientists are superior to those who are white.

Fastest Growth Consumer App

Over one million people were using ChatGPT as of December 4, 2022. ChatGPT had over 100 million users in January 2023, making it the consumer app with the fastest growth to that point.

In February 2023, OpenAI started taking signups from users in the United States for ChatGPT Plus, a premium service that would cost $20 per month. The company guaranteed access during peak hours, zero downtime, priority access to new features, and quicker response times with the improved but still “experimental” version of ChatGPT.

This enables developers to integrate ChatGPT into their applications in either an unmodified or customised form. The ChatGPT API is ten times more affordable than the GPT-3.5 models at $0.002 for 1000 tokens.

Some users were able to view the titles of other users’ discussions in March 2023 thanks to a flaw. Users couldn’t view the conversation’s substance, according to OpenAI CEO Sam Altman. When OpenAI reported that the fault had exposed customers’ “first and last name, email address, payment address, the last four digits of a credit card number, and credit card expiration date,” further reports revealed that the bug had been considerably more serious than first thought.

Guest researcher Scott Aaronson from OpenAI claims that the company is developing a tool to digitally watermark its text production algorithms to thwart spammers and other bad actors from utilising their services for academic plagiarism.

Breakthroughs Of The Year

Derek Thompson identified ChatGPT as a component of “the generative-AI eruption” that “may change our mind about how we work, how we think, and what human creativity really is” in The Atlantic magazine’s “Breakthroughs of the Year” for 2022.

These are not the types of individuals who gush over every shiny new product. Clearly, a significant event is taking place. In a tweet, Elon Musk said, “ChatGPT is terrifying excellent. AI that is dangerously powerful is not far away.”

Employees at Google reportedly put a chatbot named “Apprentice Bard” through rigorous testing before Google launched it as Google Bard, a rival to ChatGPT.

The ChatGPT scored 42%, falling short of the required pass rate of 55%.

AI Could Aid In Educational Objectives

Education professionals, academics, journalists, artists, ethicists, and public activists have all criticised ChatGPT since its release.

Reporters have made remarks about ChatGPT’s propensity to “hallucinate.” He requested ChatGPT for, for instance, “the largest country in Central America that isn’t Mexico.” When Nicaragua should have been the answer, ChatGPT responded with Guatemala.

Writing a song is “a blood and guts business that requires something of me to initiate the new and fresh idea,” the songwriter said on The Red Hand Files in response to being handed a song via ChatGPT in the vein of Nick Cave in January 2023.

According to reports, ChatGPT gained 4.9% in its dummy account with 38 equities, compared to an average loss of 0.8% for the 10 benchmarked investment funds.