Artificial intelligence: Microsoft fires team responsible for AI ethics

0
82

The wave of layoffs in the tech sector does not stop at Microsoft either. However, the team responsible for ethics and society in the field of artificial intelligence is also affected. It has to be complete, reports the Platformer newsletter.

At its peak in 2020, this team comprised around 30 employees. These included developers, designers and philosophers. Since then, however, the number of jobs has already been cut, and in October 2022 only 7 employees were left after restructuring.

AI regulations and product design

The team's task was to transfer Microsoft's AI principles into the product design. These guidelines are set by the Office for Responsible AI, which still exists. However, according to the platformer report, the AI ​​ethics team was not directly involved in the development. Rather, it sensitized the development teams and explained the risks. In this way, it should ensure that AI principles are also transferred to the products. Most recently, it identified the risks emanating from the integration of the GPT language model into Microsoft's product portfolio.

In a statement, Microsoft said it appreciated the team's “groundbreaking work.” It has made a contribution to the responsible use of AI solutions in the group. In the meantime, the number of employees responsible for the implementation of the AI ​​principles has been increased company-wide.

Microsoft is committed to building AI products and experiences safely and responsibly. This is done by investing in people, processes and partnerships that prioritize this aspect.

Statement Microsoft

In plain language this means: A dedicated team responsible for AI ethics is no longer required.

Microsoft and the AI ​​development: “Very, very high pressure”

According to information from Platformer, there is “very, very high pressure” internally at Microsoft to integrate the current GPT models from OpenAI into the products. This strategy is said to come directly from CTO Kevin Scott and CEO Satya Nadella. The result is well known: Microsoft has been fueling the AI ​​hype in recent weeks and months, and ChatGPT has been integrated into the product range at high speed. The chat search in Bing has benefited the most so far.

But the problems that Microsoft's AI chat has to deal with are also a price to pay for speed. This includes incorrect and bizarre to bizarre answers that the bot produces. To get around the difficulties, Microsoft limited the number of search queries – the number of chat turns per session is currently 10.

Background on the topic:

< ul class="text-ul text-width">

  • Large Language Models and Chatbots: The Technology Behind ChatGPT, Bing Chat and Google Bard
  • AI search engine: Microsoft presents Bing and Edge with ChatGPT integration
  • Google Bard: AI competitor to Microsoft Bing-ChatGPT has a name
  • AI language model “LLaMA”: Meta presented Competitor for ChatGPT and Google Bard
  • Internally, however, it rumbles in other places. Last year, for example, this affected the Bing Image Creator, which is based on OpenAI's Dall-E model and can generate images based on text input. The preview version started in some countries, but according to the platformer report, the product was not yet mature. Internally, there were warnings about potential consequences if generated images closely resemble artists' works. It is also still unclear how it should be evaluated when complex AI models are trained with copyrighted material. For this reason, Getty Images is already suing the AI ​​image generator Stable Diffusion.

    Launch wave is rolling over tech industry

    Layoffs are taking place across the tech industry. At Microsoft there should be a total of around 10,000 employees who have to go. Just today, Meta announced that it would lay off around 10,000 employees and not fill 5,000 vacancies.

    Elon Musk is now notorious for his clear cuts on Twitter. Other companies cutting jobs include Amazon, Google, Intel and Dell.