Home Bots & Brains ‘GPT-4 to be released this week’

‘GPT-4 to be released this week’

by Marco van der Hoeven

Microsoft Germany has announced that GPT-4 will be released this week, according to the company’s CTO, Andreas Braun. The announcement was made during an event called “AI in Focus – Digital Kickoff” last week. The event focused on large language models (LLMs) like GPT, and their potential to disrupt companies, as well as Microsoft’s Azure-OpenAI offering.

Braun mentioned that GPT-4 will offer multimodal models that will enable new possibilities, such as videos. Multimodality will make the models more comprehensive and will teach machines to understand natural language. According to Braun, LLMs are game-changers, and they will allow machines to understand what was previously only readable and understandable by humans.

Marianne Janik, the CEO of Microsoft Germany, also spoke at the event about the potential of AI to create value in companies. She emphasized that disruption does not necessarily mean job losses and that traditional job descriptions are changing, leading to the emergence of exciting new professions. Janik recommends that companies form internal “competence centers” to train employees in the use of AI and bundle ideas for projects.

Clemens Siebler, a senior AI specialist at Microsoft Germany, discussed some practical use cases for AI, such as speech-to-text telephone calls, which could be recorded and processed by AI, saving call center agents time. According to Siebler, AI could save a large Microsoft customer in the Netherlands 500 working hours a day. He also mentioned AI-assisted document processing and semi-automation by processing spoken language in the call and response center as other use cases.

When asked about operational reliability and fact fidelity, Siebler admitted that the AI will not always answer correctly, and it is necessary to validate. Microsoft is currently creating confidence metrics to address this issue. Customers often use AI support only on their own data sets, primarily for reading comprehension and querying inventory data, where the models are already quite accurate.

Overall, the event highlighted the potential of large language models and AI to disrupt traditional job descriptions and create exciting new professions. It also demonstrated the potential of AI to save companies time and resources, particularly in areas such as call centers and document processing. With the release of GPT-4, Microsoft and OpenAI will be taking multimodality to the next level, offering new possibilities for natural language understanding.

Misschien vind je deze berichten ook interessant