To say that 2023 was a big year for AI is an understatement. During the last year, the world had to begin learning how to operate in a post-AI world. With the technology growing in scale, industries, and users, there is no telling what 2024 will bring us. So let’s take a look at the top AI stories and AI news that shaped 2023 and in doing so, we may see the direction of AI moving forward.
At the start of 2023, the European Union unveiled a first-of-its-kind set of regulations aimed at artificial intelligence, which was named the AI Act. This was pushed due to questions surrounding responsible AI and AI & ethics which has been circulating in the data science community for years.
Software giant Microsoft added some incredible enhancements to its search engine Bing with a focus on artificial intelligence capabilities. This was a venture between Microsoft and ChatGPT creator OpenAI. New features such as a Chatbot, DALLE-2 image generation, and more went live after March.
Last year, Microsoft increased its investment in OpenAI by $10 billion, at a $29 billion valuation. This left the company with a 49% stake in the AI start-up. When this came to light, it was clear that the tech giant was emphasizing the potential benefits of AI integration within its ecosystem of products such as Bing, Edge, Office, and other products.
In February of last year, the world finally saw Google’s answer to ChatGPT, Bard. According to the company, it was a new AI service that promised to revolutionize conversational AI by being powered by the LaMDA model. Google & Alphabet CEO Sundar Pichai wrote in a blog post about BARD, “Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence, and creativity of our large language models.”
Not to be outdone, Meta related its own model named LLaMA. According to the company, it is a 65-billion-parameter state-of-the-art language model designed to assist AI researchers and practitioners in developing and deploying NLP models who may not have access to the infrastructure required. Unlike the GPT or LaMDA models, LLaMA is open-sourced as Meta pushes for greater AI democratization.
Though rivals Meta and Google pushed out their own models to the world, people were still excited at the prospect of GPT-4. Thanks to a Microsoft Germany CTO, the fire increased. Of the model, he said, “We will introduce GPT-4 next week, there we will have multimodal models that will offer completely different possibilities – for example, videos.” He also described it as a “game changer.”
Before the release of GPT-4, OpenAI went public with a short summary of abilities users could expect from the chatbot. Part of the summary included the release of partnerships with multiple companies such as Dulingo and Khan Academy aimed at using AI to supercharge educational aspirations.
When GPT-4 went live, it took the internet by storm and it didn’t take long for users to push the boundaries of what the AI was capable of. From creating a Doom Clone to having the AI help users with web development and coding tasks, the release of GPT-4 brought with it the desire for users to see how far they could go with AI.
Fans of the PyTorch framework got an early Christmas present with the release of PyTorch 2.0. As you can imagine, the release has been eagerly anticipated by the data science community, and from the looks of it, the team did not disappoint. As one would also expect, with the update a host of new features and improvements to the platform, making it even more powerful and versatile than before.
In a report by investment giant, Goldman Sachs, up to 300 million full-time jobs can be affected by ChatGPT-like technology. This “significant disruption” in an already uncertain labor market could have unexpected consequences and is likely one reason why researchers and technology experts have asked for a “pause” in AI development. Though many have fears of job loss, others point to the possibility of AI further helping humans with their tasks, which would lower stress overall.
CEO and co-founder of Databricks, Ali Ghodsi, took to LinkedIn to introduce to the world, Dolly 2.0 – the world’s first open-source LLM that is instruction-following and fine-tuned on a human-generated instruction dataset licensed for commercial use. Dolly 2.0 is capable of following instructions, enabling organizations to build, own, and customize LLMs for their specific needs. This means, that if a company wants to use an LLM for sentiment analysis of customer reviews, they don’t have to start from the foundations.
Stability AI introduced StableVicuna, the first large-scale open-source chatbot trained via reinforcement learning through human feedback or RLHF. According to the company, it is a further instruction fine-tuned and RLHF-trained version of Vicuna v0 13b, which is an instruction fine-tuned LLaMA 13b model.
In April, Google announced that the DeepMind AI research team and the Brain team from Google Research will merge to form a single unit called Google DeepMind. The company hoped that this move would accelerate progress in AI and develop more capable AI systems safely and responsibly.
Bard was only the beginning, that’s because in May the company introduced PaLM 2, the second generation of its large-scale language model. PaLM 2 which has been extensively trained in multilingual text, encompassing more than 100 languages. The goal of PaLM 2 was to help power the Carbon AI control system and bring human-like intelligence to Google’s products and services, including Bard, Gmail, and Google Docs.
Stability AI and its CarperAI lab unveiled Stable Beluga 1 and its successor, Stable Beluga 2 (previously known as FreeWilly). The goal of these two large language models is to expand and create a new standard for open-access AI research. According to Stability AI, Stable Beluga 1 builds upon the foundation of the LLaMA 65B model. It is fine-tuned with a new synthetically generated dataset using Supervised fine-tuning (SFT) in the standard Alpaca format.
Finally, back in the fall of 2023, Amazon announced Amazon Bedrock, a new development in generative AI that Amazon promises will reshape the landscape of artificial intelligence. Swami Sivasubramanian, vice president of Data and AI at AWS said, “Over the last year, the proliferation of data, access to scalable compute, and advancements in machine learning have led to a surge of interest in generative AI, sparking new ideas that could transform entire industries and reimagine how work gets done.”
Though OpenAI started off strong at the beginning of 2023, it wasn’t quite done and the company unveiled its latest version of DALL-E as part of a research preview, called DALL-E-3 According to the AI startup, DALL-E 3 “understands significantly more nuance and detail than our previous systems.”
Google DeepMind CEO and Co-Founder Demis Hassabis and Sundar Pichai, CEO of Google both introduced Gemini, the tech giant’s highly anticipated AI model that promises to reshape the landscape of artificial intelligence. According to the company, Gemini promises to deliver cutting-edge capabilities and may put OpenAI on alert as the model comes with three optimized configurations to better suit the needs of different users.
As OpenAI kickstarted the year, Google closed it with yet another announcement. In the middle of last month, Google Research and UC San Diego introduced PixelLLM, a sophisticated vision-language model that pioneers fine-grained localization, dense object captioning, and vision-language alignment enabling localization tasks.
Looking forward to 2024
What a year right? Who would have thought the direction of 2023 would have been so heavily affected by AI? And as we’ve seen over the last couple of years, AI is no longer a part of just data science, but spreading into every facet and industry. One can only imagine what 2024 has.
But if you want to keep up with the latest developments, trends, frameworks, and news, then you don’t want to miss ODSC East 2024. At East, you’ll come face to face with the future of AI, and the future of our world thanks to AI. So, what are you waiting for? Get your pass today!
Want to get AI news sent directly to your inbox every Friday? Sign up for our new weekly AI Weekly Recap newsletter here!