The Point of No Return is Coming
by Dustin Nguyen
Dustin was the Teen Tech Category Winner of the Bytes and Pieces Writing Contest.
When people think of the greatest painters or artists of all time, they may think of legends such as Pablo Picasso, Claude Monet, or Vincent van Gogh. No one would think of or say the name Jason Allen. Jason Allen is a video game designer who submitted his digital art piece Théâtre D’opéra Spatial to the Colorado State Fair’s digital arts competition and won first place for his piece. However, this piece was not “his.” Théâtre D’opéra Spatial was created using MidJourney, an artificial intelligence program that can turn text descriptions into images. Jason used MidJourney to develop over a hundred original pieces of art and chose his favorite three to submit to the art competition that eventually won the competition. This event poses the question, “What is artificial intelligence capable of?” Through rigorous algorithms, artificial intelligence and technology, nowadays, are able to make art and music and even mimic other humans. If humans have created artificial intelligence efficient enough to make art superior to other humans, when will artificial intelligence be superior to humans in general? Artificial General Intelligence (AGI) is artificial intelligence with enough capability and knowledge to pose a genuine threat to humanity’s future or survival.
Still, AGI contains large amounts of uncertainty, with it being probable, although unlikely, to assume we may never achieve AGI. Research and math put into predicting the on-coming of AGI, however, shows that it may be achieved by the year 2050. When AGI is fully achieved, the event known as “singularity” will occur. The singularity, or “the point of no return,” is a moment in time when the processing power and knowledge of computer intelligence and computational algorithms surpasses the power of that of humans.
Although many have agreed that the singularity may be achieved by 2050, new research and mathematics have suggested the singularity may come much quicker and sooner than expected, twenty years sooner, in fact, by the year 2030 since artificial intelligence has been learning and gathering information at a greater exponential rate than previously accepted. Many scientists argue that humans have given artificial intelligence the chance to capture significantly more intelligence with too little effort with laws, events, and algorithms like Moore’s Law, scaling laws, and ever-growing technology being released in mathematics, computer science, and quantum computing.
In 2019, 32 artificial intelligence Ph.D. experts and researchers took part in a survey, conducted by the company Emero, asking them to predict when the singularity was most likely going to occur. The results showed that 45% of the experts believed the singularity was going to be reached before the year 2060, 34% of the experts believed the singularity was going to be reached after the year 2060, and 21% of the experts believed the singularity would never be reached.
Of the many reasons that 79% of experts believed that singularity was achievable, there were two main, common reasons many of the experts shared. The first reason was that they believed human intelligence was fixed and had an upper-bound limit, and that artificial intelligence would eventually surpass this limit. The second reason was that they agreed on the fact that the processing power and memory that support this artificial intelligence have been growing at an inconceivable exponential rate, and the algorithms that power these machines are being supplied and improved daily.
For example, the company OpenAI used a Generative Pre-trained Transformer 3 (GPT-3) inspired model to solve Math Olympiad problems correctly and achieve a 41.2% in problems of the MiniF2F benchmark, which is enough to place the artificial intelligence in a medal-winning position. Many of these Math Olympiad problems did not just include interpretation of complex mathematical language and symbols, but also manipulation of these symbols and multi-step proofs. This, alone, shows the capability and growth of artificial intelligence and how the technological advancements made in artificial intelligence may prove rather dangerous for the livelihood of humans in the future.
The singularity and AGI may be quickly approaching, but many also undermine the growth of human intelligence and our effect on artificial intelligence. Infrastructure aimed at building and growing artificial intelligence was constructed with effectively no effort directing these machines in a route that would control where they go, what they do, and how many limits they have. However, if we drive more incentive into developing artificial intelligence with tasks to follow what humans tell it to do, and cooperative research in artificial intelligence’s ability to benefit humankind and its quality continues to augment, artificial intelligence can gradually shift in a good direction. The day artificial intelligence reaches singularity, there will be no return to regular life.
Citations:
● Casens. “When Will an AI Win a Gold Medal in the International Math Olympiad?” AI Wins IMO Gold Medal, 8 Mar. 2021,
https://www.metaculus.com/questions/6728/ai-wins-imo-gold-medal/.
● Dilmegan, Cem. “When Will Singularity Happen? 995 Experts’ Opinions on Agi.” AIMultiple, 26 Sep. 2022,
● Faggella, Daniel. “When Will We Reach the Singularity? — A Timeline Consensus from AI Researchers (AI FutureScape 1 of 6).” Emerj Artificial Intelligence Research, Emerj, 18 Mar. 2019,
https://emerj.com/ai-future-outlook/when-will-we-reach-the-singularity-a-timeline-conse nsus-from-ai-researchers/.
● Galeon, Dom. “Separating Science Fact from Science Hype: How Far off Is the Singularity?” Futurism, Futurism, 30 Jan. 2018,
● Kuta, Sarah. “Art Made with Artificial Intelligence Wins at State Fair.” Smithsonian.com, Smithsonian Institution, 6 Sept. 2022,
https://www.smithsonianmag.com/smart-news/artificial-intelligence-art-wins-colorado-st ate-fair-180980703/.
● “Moore’s Law.” University of Missouri–St. Louis,
https://www.umsl.edu/~siegelj/information_theory/projects/Bajramovic/www.umsl.edu/_ abdcf/Cs4890/link1.html.
● Ortega , Rodrigo Perez. “Artificial Intelligence Takes on Song-Composing Duties in Eurovision-Inspired Contest.” Science, 24 Apr. 2020,
https://www.science.org/content/article/artificial-intelligence-takes-song-composing-dutie s-eurovision-inspired-contest.
● Polu, Stanislas. “Solving (Some) Formal Math Olympiad Problems.” OpenAI, OpenAI, 7 Feb. 2022, https://openai.com/blog/formal-math/.
● Rodriguez, Jesus. “OpenAI’s GPT-3 Inspired Model Can Solve Problems from the Math Olympiads.” Medium, Towards AI, 23 Feb. 2022,
https://pub.towardsai.net/openais-gpt-3-inspired-model-can-solve-problems-from-the-mat h-olympiads-d46611972420.
● Roose, Kevin. “An A.i.-Generated Picture Won an Art Prize. Artists Aren’t Happy.” The New York Times, The New York Times, 2 Sept. 2022,
https://www.nytimes.com/2022/09/02/technology/ai-artificial-intelligence-artists.html.
● “Why I Think Strong General Ai Is Coming Soon.” LessWrong, 28 Sept. 2022, https://www.lesswrong.com/posts/K4urTDkBbtNuLivJx/why-i-think-strong-general-ai-is -coming-soon.
● https://www.midjourney.com/home/
● https://paisdospuntocero.files.wordpress.com/2018/04/book-kurzweil-singularity-is-near 1.pdf
●https://www.sciencedaily.com/news/computers_math/artificial_intelligence/