Shulou(Shulou.com)11/24 Report--

January 22, AIGC (generative artificial Intelligence), as free online AI chat tools, such as ChatGPT and Dall-E 2, is becoming so popular that anyone can use them to generate text or art images. Now, Microsoft is making a big bet on AI, which could help chat robots such as ChatGPT become more popular.

Tu Yuan Pexels for years, Microsoft has been catching up with Google in the AI competition, and the software giant hopes to take a step forward through a large investment in OpenAI. OpenAI is committed to developing AI models that can not only help people format email, but also analyze Excel spreadsheets, generate AI images to help enrich PowerPoint presentations, and even write complete emails in Outlook. And this is just the beginning.

A few days ago, Microsoft began to introduce OpenAI services into the Azure cloud platform, providing enterprises and startups with the ability to integrate models such as ChatGPT into their own systems. Microsoft has built AI tools into many of its consumer products, such as the Dall-E 2 feature in its Bing search engine, to generate images based on text prompts. Recently, there have been media reports that Microsoft is trying to put more AI tools into Microsoft Office.

"eventually, every Microsoft product will have some of the same AI features," Satya Satya Nadella, Microsoft's chief executive, said at the Davos forum this week.

Microsoft is promoting its claim that the technology industry will revolutionize AI. Microsoft and several companies, including rival Google, believe that their big bet on AI will change not only productivity software, but also the entire technology industry, thanks to so-called large language models. These models can understand human language, talk to human beings, and imitate human beings in various fields such as writing, artistic creation, computer programming and so on.

Referring to the latest advances in large language models, Bojan Tunguz, a Nvidia machine learning modeler and data scientist, said: "I think this will fundamentally change the future of all knowledge work. Of course, we are only seeing the early stages of this trend."

This is an optimistic view, which is often heard in Silicon Valley and Seattle these days. While Google, Microsoft and other tech giants have been testing and training AI models for years, a group of new startups have taken the lead in bringing their experimental AI programs directly to the public. Over the past year, the release of tools such as ChatGPT, Dalle-2, Stability AI's Stable Diffusion and Midjourney has made generative AI the hottest buzzword in the industry.

Now, big technology companies are catching up. Among them, Microsoft seems to invest the most money and resources. At Davos, Nadella predicted that the current AI would trigger an industry-wide "platform shift", just as it has moved to mobile devices and cloud computing platforms over the past 15 years. Microsoft is understood to be one of the early investors in OpenAI, which plans to buy more shares in OpenAI for $10 billion. Microsoft and OpenAI declined to comment.

Some people are skeptical that generative AI is just a dazzling toy and novelty, but it is mediocre in practical application, and may even have more negative effects.

ChatGPT can generate plausible texts in many areas, but it is prone to factual errors and biases. Many school districts in the United States have banned teachers and students from using ChatGPT because it may be used as a cheating tool. Tech news site CNET has been criticized for quietly using AI to write articles because some articles have been found to have errors.

Stability AI, the developer of AI art generator Stable Diffusion, has been sued by Getty Images for allegedly using its copyrighted work to train AI models without permission. Microsoft has also made mistakes in developing the AI model, including the chat robot Tay released in 2016, which made remarks such as racial discrimination.

These risks help explain why Google has developed many very advanced AI chat tools that are not yet available to the public. Google's chatbot system, LaMDA, is so complex that one of the company's engineers even believes it has perceptual capabilities, reigniting the debate over whether it is irresponsible to provide similar tools to ordinary users.

Recently, Yann LeCun, Meta's chief AI scientist, told a forum hosted by Collective [I]: "some people will get hurt, which will be inevitable, but this should not be an obstacle to AI's progress. However, for companies involved in the development of new AI, it is essential to find ways to mitigate the harm."

Mr Le Kun believes that ChatGPT is "doing well, but not revolutionary". He explains that ChatGPT comes from startups rather than tech giants because "both Google and Meta suffer huge losses by launching systems that make things up". In November last year, Meta released a language model called Galactica for scientists, but just three days later, the model was stopped because of strong opposition.

Still, Microsoft's bet on OpenAI has put new pressure on its competitors, especially in the lucrative cloud computing space. While using ChatGPT to improve its products can help Microsoft maintain its advantage in productivity software, the fiercer competition is to sell AI services to enterprises. These may include established companies that want to build smarter customer service chat robots, startups that develop more specialized AI tools, and even other AI companies that need cloud computing power to train their models.

One application that has received attention is the use of AI to help software developers write code. GitHub, a subsidiary of Microsoft, uses OpenAI technology in a tool called GitHub Copilot, which provides code advice in real time as programmers program.

Startup Neeva recently launched a search engine based on AI support, which uses a language model to answer users' questions directly. "Microsoft's adoption of OpenAI technology will definitely lead to wider use of it," said Sridhar Ramaswamy Ramaswami, chief executive. "the problem is that once Microsoft starts to put pressure on a particular area, they tend to crowd out everyone else."

Ramaswami predicts that Google may release its own chat robot next month, "and it will be very good." But he believes it remains to be seen whether Microsoft or Google can monopolize the AI chatbot market in the long run. "this technology is rapidly commoditising," says Ramaswami, whose start-ups like Anthropic and Cohere are already developing their own tools and models. "startups can afford to take the risk because they have much less to lose. I think this will be a challenge for Google and Microsoft."

Sandal Pichai (Sundar Pichai), chief executive of Google's parent company, which announced 12000 layoffs on Friday, told the rest of its employees that the company would pay more attention to AI. Not long ago, Jeff Dean, head of Google's AI business, published a blog post highlighting the company's latest developments in large language models, and Google is said to have called on its founders, Larry Page and Sergey Brin, to review their AI strategy.

At the same time, Microsoft's big bet on AI has created an exciting atmosphere for the 47-year-old company at a time when the technology industry is in the doldrums. Microsoft recently announced its own layoffs after several rounds of layoffs at Amazon, Meta and others, but its big bet that AI may give its remaining employees and future new employees more reason to be optimistic.

To deliver on these promises, Microsoft must demonstrate its ability to transform its new AI tools into a more important asset than Clippy 2.0.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.