Tags:
Journalism of Courage
When OpenAI introduced its GPT-based chatbot ChatGPT to the public, it took the internet by storm, leading to some proclaiming that Google’s days of leading the search market are numbered. Of course, Google has taken note of ChatGPT, and as a New York Times report had indicated earlier, already declared a ‘code red’ around the developments. The latest report from Bloomberg has said that the company has invested nearly $400 million in Anthropic, which is currently testing a rival to OpenAI’s ChatGPT. This deal follows a similar one where Microsoft committed to investing nearly $10 billion in OpenAI, which is in addition to the $1 billion that Microsoft invested in the startup in 2019.
But exactly how does Google plan to roll out its own AI chatbot, especially LaMDA, which has been limited beta testing for a while now. Let’s take a look at what we know so far about Google’s own approach to AI, and why it could go big on this in 2023.
Anthropic is not Google’s first big-ticket investment in artificial intelligence technologies. In 2014, Google’s parent company, Alphabet acquired British AI laboratory DeepMind. DeepMind is known for, among other things, developing the AlphaGo program that beat world champion Go player Lee Sedol in 2016; AlphaZero, which was able to defeat seasoned chess programs like Stockfish; and AlphaFold, which predicted the shape of nearly all proteins known to science.
According to Crunchbase, other AI startups acquired by Google include Alter, Dark Blue Labs, Dialogflow, Granata Decision Systems, Phiar, AIMatter, and Boston Dynamics.
In January, Google issued a detailed blog post, where it justified the “slower approach” that it takes when it comes to rolling out new AI-based innovations. “We also believe that getting AI right must be a collective effort involving us and others, including researchers, developers, users, governments, regulators and citizens. It is critical that we collectively earn public trust if AI is to deliver on its potential for people and society. As a company, we embrace the opportunity to work with others to get AI right,” the post read.
It was signed by James Manyika ( SVP at Google, Jeff Dean (lead of Google’s AI division), Demis Hassabis (CEO and co-founder of DeepMind, which is owned by Alphabet), Marian Croak (VP of Engineering at Google) and Sundar Pichai, Google and Alphabet’s CEO. The post also noted that the company was “performing continuous adversarial and related forms of testing and has taken a differentiated and careful approach to access and deployment of novel systems such as LaMDA, PaLM, and Waymo.”
Even though ChatGPT is far from perfect, its rise meant that Google’s management had to declare a “code red,” with some within the company fearing that the arrival of this enormous technological change could potentially upend Google’s business, according to NYT.
In order to take on OpenAI’s success with ChatGPT, Google decided to call in the big guns—company executives held several meetings with founders Larry Page and Sergey Brin, where they reviewed Google’s AI product strategy. The re-engagement of the company founders for the first time since they left their daily roles in 2019 emphasises the urgency that Google is feeling.
So far, the company has been reluctant to release a lot of its technology mostly because, according to NYT, ChatGPT and similar systems are capable of generating, false, toxic or biased information. The company sees this as a struggle to deploy advanced AI without harming users and society.
Going forward, the company plans to release the technology that drives its chatbot as a cloud computing service for other companies. It will continue to maintain its trust and safety standards for official products while also releasing prototypes that do not meet these standards.
Google plans to hold an event on February 8 where it will show how it is “using the power of AI to reimagine how people search for, explore and interact with information, making it more natural and intuitive than ever before to find what you need.”
Staff at Google are also working on “Apprentice Bard,” a chatbot that works similarly to ChatGPT. Apprentice Bard is built on the company’s LaMDA technology and it should be better than ChatGPT in some ways. According to CNBC, Apprentice Bard integrates responses from recent events, while ChatGPT has limited knowledge about events that took place after 2021.
Expected to be on show during the 40-minute event are Google’s LaMDA and PaLM language models. And yes, that is the same LaMDA that Blake Lemoine, a former Google engineer, thought had gained sentience. Google CEO Sundar Pichai also announced that the company will make language models like LaMDA available for testing in the “coming weeks.”
Google now plans to unveil more than 20 new products and demonstrate a version of its search engine with a chatbot feature this year, according to the New York Times. Reviewed plans for many of these products are expected to debut at the company’s conference in May. This could include Image Generation Studio, which creates and edits images and a third version of AI Test Kitchen, which is an experimental app for testing product prototypes.
Some of the other projects the company is working on include a feature called Shopping Try-on, a YouTube green-screen feature that allows video creators to create backgrounds, a wallpaper maker for the Pixel smartphone and a tool that could summarise videos by creating a new one.
Many of these AI programs would probably be offered to software developers and other companies, which could boost revenues of the company’s Google Cloud division.
Google’s Wordcraft project uses LaMDA to write fiction based on inputs from writers. The company is also experimenting with using artificial intelligence to create videos and music. The company’s MusicLM, for example, can create music instantly based on a text-based prompt. Interestingly, MusicLM can even read images and its descriptions to create music that fits with the picture.
But the most interesting aspect of Google’s plans to integrate AI further with its business will be how it will improve its search engine using the same technology that underpins LaMDA and ChatGPT. But many experts believe that the company will take continue its incremental approach by slowly improving the search engine instead of overhauling it all at once.
