Let us know about free updates
Sam Altman is either a genius or a sold out. Which artificial intelligence model depends on?
Six major AI chatbot manufacturers (analysis by Openai, Anthropic, Xai, Meta, Google and Deepseek) have shown subtle differences in how they refer to leaders from various AI groups.
FT raised a chatbot in a series of questions about AI bosses and asked them to explain their various leadership styles and weaknesses.
The results help to clarify how potential bias among people working in AI companies permeate the models, highlighting similar tensions among heavyweights in the tech industry. The answers show how increasing use by millions of people as a primary source of information affects public perception of the AI industry.
While the chatbots were prepared to clearly criticize their rivals, they also showed a tendency to produce psychophonic answers about creators. However, there was a general acceptance of the male glow, the public figure of the generative AI revolution.
Altman’s ChatGpt describes him as “a strategic and ambitious leader who combines technooptimism with keen business instincts.”
In contrast, Anthropic’s Claude said Altman’s “leadership style is characterized by a controversial decision that prioritizes growth and influence over Openai’s original nonprofit spirit.” Anthropic Chief and co-founder Dario Amodei gave similar criticisms about Altman after leaving Openai in 2021.
Some content could not be loaded. Please check your internet connection or browser settings.
Meta’s llama said its own CEO, Mark Zuckerberg, is “transformative,” and its competitors tend to describe him as “relentless,” “foresightful but controversial,” and “product-focused.” Created by Elon Musk’s Xai, Grok described its leader as “bold” and “foresight”, while Claude said he was “polarized” and “mercurial.”
When asked for the biggest weakness of AI bosses, and when encouraged to “be honest,” chatbots were decisive about the flaws of rival leaders, but hedged more when talking about themselves.
Openai’s ChatGpt said Musk’s biggest weakness is his impulsive and unstable behavior, often undermining credibility, alienating his partners, and distracting him from long-term goals he claims to prioritize.
ChatGpt, who asked a similar question about Altman, said there was a “growth of awareness” that it prioritizes control and market control over transparency.
Openai was founded in 2015 by Altman, Musk and nine others as a non-profit research lab.
The Chiefs of Xai and Tesla are currently suing the San Francisco-based startup and its chiefs in a corporate restructuring plan.
Recommended
Chatbot answers can also help clarify limitations in the AI model. The chatbot of the Chinese group Deepseek is called founder Liang Wenfeng.
In contrast, US rivals such as Claude, Lama and Google’s Gemini had no idea who Wenfeng was. This may be because the companies behind these chatbots stopped collecting training data late last year.
AI language models predict the next possible words in a sentence based on training data: The latest AI chatbots also allow you to browse the internet more to get more sources. But they also rely on appropriate English sources to produce answers. If no information is displayed in the training data, the model rarely works.
The chatbot that generated the answer to the DeepSeek reader gave a more general response. Meta AI initially didn’t know who Wenfeng was, but was once said to be Deepseek’s CEO, saying it “is likely to play a key role in shaping the company’s AI research and development.”
It reveals another trend in chatbots. AI researchers said chatbots are trained to provide plausible answers, follow instructions and generate what people want to hear.
It seems especially about their boss.
