Dec, 28, 2022.
Artificial intelligence has become integrated into every aspect of our lives. From chatbots and virtual assistants to automated industrial machinery and self-driving cars, it’s hard to ignore its impact. Artificial intelligence is an indispensable companion for many people and businesses in various areas of life and what makes it even more vital is the constant innovations that push the boundaries of what this technology can do.
Artificial intelligence has been evolving at an accelerating rate over the past few years. The pace of AI news in 2022 has been unrelenting and quick; as soon as you understood how things stood in AI, a new study or breakthrough would render that knowledge obsolete.
When it comes to generative AI, which can create art made up of text, images, audio, and video, we will probably hit the knee of the curve in 2022. After years of research and development, deep-learning AI finally made it into commercial applications, enabling millions of people to experience the technology for the first time. AI innovations have raised eyebrows, sparked debates, provoked existential crises, and inspired astonishment.
This article will introduce you to a list of the biggest AI news and trends of 2022.
ChatGPT Speaks to The World
ChatGPT is a prototype dialogue-based AI chatbot capable of understanding natural human language and generating impressively detailed human-like written text. It is the latest evolution of the GPT – or Generative Pre-Trained Transformer – family of text-generating AIs. In an instant chat, users can ask ChatGPT questions, and the AI will respond in full sentences while attempting to imitate the flow of a conversation. However, its responses are not always suitable or right.
In November, OpenAI launched ChatGPT, a chatbot powered by their GPT-3 big language model. In order to gather information and suggestions from users on how to improve the model and deliver more accurate and potentially less detrimental outcomes, OpenAI made it freely accessible on its website.
Sam Altman, CEO of OpenAI, tweeted five days after ChatGPT's release that it had gained over a million users. People used it to make recipes, write poetry, aid with programming tasks, and much more. Additionally, users discovered a way to circumvent prohibitions against the tool and make it respond to potentially dangerous questions using prompt injection attacks.
Although OpenAI's ChatGPT harnessed the best of what GPT-3 had already been offering since 2020 (with some substantial upgrades behind the hood), The free price tag ensured that ChatGPT was the first time a large audience had experienced what OpenAI's GPT technology can achieve.
DALL-E 2 Astonishing Artwork
DALL-E 2, a deep-learning image-synthesis model that stunned people with its incredible capacity to produce images from text prompts, was unveiled by OpenAI in April this year. Thanks to a method known as latent diffusion, DALL-E 2 was trained on hundreds of millions of images from the Internet and was able to create unique imagery combinations.
When version 1 of DALL-E was released, it struggled to render low-resolution images. Version 2 of DALL-E illustrated fascinating artistic works at 1024×1024 resolution, and before you knew it, Twitter was soon filled with pictures of astonishing photorealistic creative work.
Due to concerns about abuse, OpenAI initially restricted the use of DALL-E 2 to 200 testers. Content filters prohibited sexual and violent prompts. DALL-E 2 was eventually made accessible to everyone in late September after OpenAI gradually accepted over a million people to a closed trial.
Google engineer "Blake Lemoine" Believes LaMDA is Sentient
The Washington Post reported in early July that a Google engineer called Blake Lemoine had been placed on paid leave because he thought Google's LaMDA (Language Model for Dialogue Applications) was sentient and deserving of equal rights to those of a human.
Lemoine started talking to LaMDA about religion and philosophy while working for Google's Responsible AI group and claimed he could tell the text was truly intelligent. Lemoine said to the Post, "It doesn't matter whether they have a brain made of meat in their head. Or if they have a billion lines of code. I talk to them. And I hear what they have to say, and that is how I decide what is and isn't a person."
Google responded that LaMDA was not actually sentient and was merely telling Lemoine what he wanted to hear. LaMDA had been trained on millions of books and websites, just like the text generation tool GPT-3. It predicted the most likely words to be said in response to Lemoine's input. Lemoine was let go by Google later in July for breaking their data security policies. He wasn't the only one in 2022 to fall for the hype surrounding an AI's extensive language model.
Stable Diffusion Open-source Image Generator
An image synthesis model comparable to OpenAI's DALL-E 2 called Stable Diffusion 1.4 was released by Stability AI and CompVis on August 22. Stable Diffusion emerged as an open-source project with source code and checkpoint files, whereas DALL-E was introduced as a closed model with substantial restrictions. (The training data for the model was processed on the cloud at a cost of $600,000). Due to its openness, any synthetic content might be created without restriction. Additionally, Stable Diffusion may be used locally and privately on PCs with a good enough GPU, unlike DALL-E 2.
Not everyone applauded Stability AI's approach as a technological accomplishment. The software's ability to produce sexual content, non-consensual pornography, alternate histories, and political misinformation was criticized by critics. Artists argued that it might steal working artists' style and force them out of business. The methods used to construct the images dataset for the model became problematic when someone learned that her private medical photos had been taken from the web without consent and a way to have them removed. Bias in the dataset used to train the model also attracted criticism.
At the same time, a few hobbyists embraced Stable Diffusion wholeheartedly and soon created an open-source ecosystem around it.
DeepMind AlphaFold Predicts Nearly All Protein Structures
DeepMind stated that its AlphaFold AI model has accurately predicted the patterns of nearly all known proteins from virtually every organism on Earth that had a sequenced genome In July. AlphaFold was first introduced in the summer of 2021 and had previously predicted the form of every human protein. But a year later, its database of protein structures had grown to almost 200 million.
To enable researchers from all over the world to access and use the data for research related to medicine and biological science, DeepMind made these predicted protein structures accessible in a public database hosted by the European Bioinformatics Institute at the European Molecular Biology Laboratory (EMBL-EBI).
AI Art wins a State Fair Art Competition
Three AI-generated images were submitted to the Colorado State Fair's fine arts category in early August by Jason Allen, a native of Colorado. He revealed before the end of the month that Théâtre d'Opéra Spatial had taken first place in the Digital Arts/Digitally Manipulated Photography competition. People were astonished when word of the win spread.
Allen used Midjourney, a commercial image-generating model that uses a customized Discord server and is comparable to Stable Diffusion but has a unique visual art style. He canvassed the three prints and presented them in the competition. A heated discussion on the nature of art and what it means to be an artist erupted on social media in response to AI's symbolic victory over humanity.
In a related manner, a significant cultural debate on the morality of AI-generated art has recently erupted. Computer scientists who created it see AI image synthesis as a necessary and positive technological step. However, artists who have trained for decades perceive it as an existential threat. People have exchanged death threats on social media, and artist communities have complained about or protested AI work. That argument is still being debated and might not be resolved anytime soon.
Meta’s CICERO Masters Diplomacy
In November, Meta unveiled Cicero, an AI model that can outperform humans in online Diplomacy games on webDiplomacy.net. That's a significant accomplishment considering that diplomacy is mainly a social game that needs intensive persuasion, collaboration, and negotiation with other players in order to succeed. Meta essentially created a model that could trick people into thinking they were playing with another person.
Meta trained Cicero's extensive language model component on text extracted from the Internet as well as transcripts of 40,000 human-played Diplomacy games from the website webDiplomacy.net to hone its negotiating abilities. In the meantime, Meta also created a strategic element that could assess the game's situation, foresee how the other players would act, and then act accordingly.
Meta believes it can utilize Cicero's lessons to power a new generation of video games with more intelligent NPCs and to lessen communication barriers between humans and AI during multi-session conversations.