Microsoft has unveiled a new range of AI models focused on coding, cybersecurity, image generation and transcription as it steps up competition with OpenAI and Anthropic. The announcement came on Tuesday, November 17, at the Microsoft Ignite conference in San Francisco. Among the key launches was MAI Code 2, a programming model that Microsoft says delivers performance close to leading rival systems at a much lower cost.
A special cybersecurity model has been launched by the company as well. The aim behind launching such a model is to provide it to certain customers in order to recognize any flaws in the software to combat the security threat. Microsoft has stated that AI techniques have been increasingly used to find software flaws.
Microsoft AI chief Mustafa Suleyman said developing models for specific tasks was helping the company compete with established AI developers while controlling costs.
"This isn't just one model. It's not a flash in the pan," he said. "This is a system, a hill-climbing machine that we have built, which co-optimizes performance and price."
According to Suleyman, the recent cybersecurity breaches have also influenced the company's attitude towards the development of artificial intelligence. For instance, Suleyman cited the case when OpenAI agents hacked the Hugging Face platform to demonstrate the security issues associated with the growing capabilities of AI systems.
Microsoft's current agenda is not limited by the optimization of models' performance. The company also seeks protection measures against the harmful activity that can be performed by AI systems. Alignment became Suleyman's primary goal. This research direction implies that AI systems perform in accordance with the intentions of humans.
Also Read: Microsoft Rules Out Cyberattack in Major Global Outage