OpenAI has recently announced its new AI ‘GPT4’ that has the capabilities of video generation, solving difficult problems with accuracy. Microsoft’s CTO had already confirmed that the firm’s new AI ‘GPT4’ is coming this week.
OpenAI’s ChatGPT had stirred up the AI and the tech companies, insisting the capabilities and beneficiaries of the new form of AI – generative AI. Since its launch, Google, Microsoft, Baidu and more got in the row in implementing artificial intelligence into their applications and services, for making them better in experience and convenience.
In the go, Microsoft’s collaboration with OpenAI now steps ahead with the announcement of GPT-4, an advanced version of GPT-3.5 that powers ChatGPT.
“We will present GPT-4 next week, because we have multimodal models that will offer completely different possibilities – for example videos,” starts Dr. Andreas Braun, CTO of Microsoft Germany. Now, OpenAI launched GPT-4, its most advanced system till date with greater problem solving efficiencies, quick and fast than ChatGPT.
Calling the large language models as game changers, CTO describes that it’s been possible to teach machines to understand natural language, as people. With the help of AI, machines can now understand languages in a statistical way what was previously readable and understandable only for people. Say, “you can ask in German and get the answer in Italian.”
GPT-4 allows for multimodal capabilities, intimating that it could put video creations into ease. Microsoft names the multimodal AI as “Kosmos-1”.
Kosmos-1 on pre-trained results show that the AI could classify images, answer questions about image content, automated labeling of images, optical text recognition and language generation tasks. Though, the modal is only 22% correct currently, the researchers are pushing it to unleash the benefits.
If released, this feat of AI would enhance and enrich the image-related searches and generative AI models.
What’s new about GPT-4?
- OpenAI’s ever advanced GPT-4 is more creative and collaborative that it can generate, edit and iterate with users on creative and technical writing tasks, like writing screeplays, learning a user’s writing style or even composing songs.
- Fascinatingly, GPT-4 answers to queries in not just text form, but also as images. Users can now ask the AI any details, captions, classifications and analyses in the image format too. For example, you can ask “What can I make with these ingredients?” with a photo displaying eggs, milk and more.
- The new GPT-4 is capable of handling over 25,000 words of text, allowing for use cases like long form content creation, extended conversations, document search and analysis.
- GPT-4 surpasses ChatGPT in its advanced reasoning capabilities, being able to interpret text language, much precisely than its predecessor.
- Assessment of the new AI reveals that GPT-4 secured higher approximate percentiles – 90 in Uniform Bar Exam and 99 in Biology Olympiad, outperforming ChatGPT, which scored 10 & 31 in the respective exams.
Applications of GPT-4
Speaking about the multimodal AI, Holger Kenn, who is the Chief Technologist Business Development AI, Microsoft trained Kosmos-1 for deploying it into Microsoft products with one of the examples including mapping millions of requests into the APIs via the cloud.
OpenAI had already collaborating with organizations and companies around the globe such as Duolingo, Morgan Stanley, Khan Academy, by building innovative GPT-4 powered platforms.
In a practically tested example, Speech-to-text phone calls in a call-center could be recorded and the agents would no longer have to manually summarize and type in the content, as AI would do the task. According to Clemens Siebler, 500 working hours could be saved daily with a large Microsoft customer in the Netherlands who receives 30,000 calls a day.
AI-supported document processing, semi-automation through the processing of spoken language in the call and contact center are the foremost beneficiary by equipping this advanced GPT-4. Means, GPT-4 would enhance the company’s exposure in B2B marketing.
However, researchers admit that AI would not always answer correctly, so it was necessary for validation. Microsoft is currently creating confidence metrics to address this issue.
“We are building a feedback loop around it with our thumbs up and thumbs down,” says AI researcher who worked on the project.
Do you think this new generative-AI can proffer mankind to lead a better lifestyle? Drop your thoughts on the AI…