With its growing popularity among the tech fraternity and increasing VC interest with billions being pumped into applications and research on generative AI, it seems 2023 is likely to witness more developments in the domain.
In November last year, San Francisco-based AI start-up OpenAI created waves after it launched its revolutionary ChatGPT.(Chat Generative Pre-trained Transformer).
Built on OpenAI’s large language models, it was opened up for public testing on November 30, 2022, and in less than a week, the application clocked 13 million users per day and earned over 100 million users in two months, making it the only internet application to do so in history.
The Indian Express shed light on the generative AI and explained as to whey it keeps Google, Microsoft & Meta on their toes.
ChatGPT took over the internet nearly a year after OpenAI introduced DALL-E to the world. DALL-E was able to create digital images from natural language descriptions or text prompts from users. There’s more, OpenAI’s other tools such as GPT-3 use deep learning and produce human-like texts to continue from prompts in natural language, while OpenAI Codex is a model that converts prompts in natural language into programming code.
The apparent question here is what is the underlying framework that is common between ChatGPT, DALL-E, Codex, and GPT-3? The answer is that they all are different facets of generative AI.
In essence, generative AI lets users create content from scratch such as texts, images, poetry, videos, audio, etc. Tools such as ChatGPT, DALL-E, and Lensa are part of generative AI, a briskly progressing sub-domain in the field of AI.
Be it the ‘Magic Avatar’ from Lensa, a feature that lets people use AI to create their stylised portraits based on their photos, or ChatGPT which offers humanish responses to text prompts, AI tools are impressing users worldwide. Images, texts, videos and audios created by AI seem to be used by one and all, and more users are drawn towards them to try their hand. All this is possible thanks to generative AI.
Until 2022, the purpose of existing AI was to analyse data, spot anomalies, detect fraud and perhaps, make recommendations such as movies to watch or best holiday destinations. Earlier AI models did this by analysing large chunks of data that it consumed. However, with generative AI, users can witness brand new content being created from scratch.
Companies around the world have realised the immense potential of generative AI and this burgeoning increase in interest can be attested to the $ 2.1 billion funding that generative AI companies received in 2022, as per PitchBook. The ability to leverage AI technologies to create entirely new content seems to be the driving force behind many tech giants that are pumping billions of dollars into generative AI.
Here is a look at how these companies are harnessing the possibilities of generative AI and deploying the new tech:
After the roaring success of OpenAI’s ChatGPT, Alphabet Inc’s Google was one of the biggest names to scramble to introduce its own bot to rival the revolutionary chatbot. The fast-paced adoption of ChatGPT seemingly threatened Google’s position as the most used search engine. Earlier this month, Google unveiled its AI chatbot Bard AI to compete with ChatGPT. Bard uses natural language processing (NLP) and machine learning algorithms which help it generate human-like responses.
Interestingly, the company has been doing research on generative AI for a long time. Its large language model LaMDA is reportedly being integrated into Google products. Google’s Pathways Language Model (PaLM), seen as a breakthrough in AI, has been trained with the Pathways system that enables it to generalise tasks across domains. Reportedly, it can be scaled up to 540 billion parameters. This means that with each task, its performance increases. On the other hand, GPT-3 only has 175 billion parameters.
Another significant development in this direction was Google’s $300 million investment in AI startup Anthropic. This deal made Google one of the biggest names to heavily invest in companies fostering generative AI.
Microsoft
In January this year, it was learnt that the tech behemoth had invested a whopping $10 billion in OpenAI, the creator of ChatGPT and DALL-E. It is the biggest investment made by the company in recent years. The massive investment follows Microsoft’s earlier investment of $1 billion in 2019 and another investment in 2021. The investment is expected to give Microsoft access to some of the most advanced AI systems in its bid to outdo its peers such as Google, Amazon, and Meta.
On February 8, the company showcased how generative AI can enhance the search experience. Microsoft introduced its brand new version of Bing Search and Microsoft Edge, both powered by ChatGP. This was a follow up to the company’s introduction of the AI-powered chatbot in its Teams Premium. The chatbot is expected to integrate meeting notes, recommend tasks, and assist in creating meeting templates for users. The company has announced that it will add ChatGPT into all its products.
Meta
The Chief Product Officer at Meta Platforms, Chris Cox, during his interaction with the media at Davos 2023, stated that generative AI has interesting applications for social media. Cox said that such technologies can help in creating image filters for its portal Instagram.
Meta has been working on generative AI research for some time. In September last year, Meta announced its Make-A-Video, a new AI system that allowed people to turn their text prompts into brief high-quality videos. The tool follows Meta’s Make-A-Scene, a multimodal generative AI that offers users more control over their AI-generated content. Meta claimed that its Make-A-Video was a testament to Meta AI’s progress in the realm of generative technology and has the potential to open new possibilities for creators and artists.
Even as Meta continues its research in this domain, there have been some misses too. In November 2022, Meta introduced its scientific knowledge-based chatbot Galactica. However, the chatbot was taken down within three days owing to misinformation and hate speech.
Apple
Apple, one of the biggest tech companies, has been conspicuous with its absence in the generative AI arms race. Interestingly, the company which revolutionised virtual assistance with Siri is no stranger to integrating AI and machine learning into its tech. While the Internet is abuzz with speculations, Apple has not revealed any of its plans around generative AI.
Amidst the Google vs Microsoft pandemonium, Apple seems to have quietly announced an employee-only event around generative AI in the second week of February. Although no official confirmation from the brand, numerous reports seem to suggest that Apple is working on AI-related applications for its products and services. A recent report suggested that Apple, although announced a hiring freeze, currently has 342 jobs open in AI-related domains such as natural language processing and deep learning. While Microsoft and Google are integrating generative AI in search, Apple is likely to bring something unique when it comes to use cases.
Baidu
Chinese search giant Baidu had earlier this year announced that generative AI will be among its prime focus going forward. The company recently announced its plans to launch its AI chatbot after internal testing. The chatbot is named ERNIE Bot. According to the company, the bot is based on Enhanced Representation through Knowledge Integration (ERNIE) which was proposed sometime in 2019. Reportedly, ERNIE can perform a wide range of tasks, including language comprehension, language generation, and text-to-image generation. “What sets ERNIE apart from other language models is its ability to integrate extensive knowledge with massive data, resulting in exceptional understanding and generation capabilities,” a Baidu spokesperson was quoted as saying.