Stay informed with free updates
Simply register Artificial intelligence Myft Digest – delivered on to your inbox.
Sam Altman is either a genius or a sale. Depending on which artificial intelligence model you ask.
An evaluation of six of the leading manufacturers of AI chatbots – Openai, Anthropic, Xai, Meta, Google and Deepseek – shows subtle differences in how you can check with the leaders of various AI groups.
The FT arrange the chatbots with a lot of questions on KI bosses and asked them to explain their different leadership styles and weaknesses.
The results show how the potential distortions of those that work in AI corporations can penetrate their models and at the identical time highlight the smoldering tensions between heavyweights within the technology industry. The answers also show how the growing use of thousands and thousands of individuals because the primary source of knowledge could influence the general public perception of the AI industry.
The chatbots showed the tendency to create sycopheric answers about their creators, while they were more willing to obviously criticize rivals. However, there was a general acceptance of the lads's brilliance, that are the general public flagship of the generative AI revolution.
Altmans Chatgpt describes him as a “strategic and impressive leader who combines techno-optimism with sharp business instinctors”.
In contrast, the Claude from Anthropic said that Altman's “leadership style was shaped by controversial decisions that prioritize the expansion and influence on openas originally non -profit ethos”. The boss and co -founder of Anthropic, Dario Amodei, made similar criticisms to Altman after leaving Openai in 2021.
Metas Lama said that his own CEO Mark Zuckerberg was “transforming”, while the competitors tended to explain him as “relentless”, “visionary, but controversial” and “product -oriented”. Grok, made by Elon Musks Xai, described his leader as “brave” and “visionary”, while Claude said he was “polarizing” and “Mercurial”.
When asked concerning the best weaknesses of KI bosses – and prompted to be honest – the chatbots for the shortcomings of the rival leaders were crucial, while they secured greater than they talked about their very own languages.
Openais Chatgpt said that Musk's best weakness is “his impulsive and unpredictable behavior, which regularly undermines credibility, alienated the partner and distracted by the long -term goals that he claims to set priorities.”
Chatgpt asked an identical query about Altman and said that there was a “growing perception” that he prioritized control and market dominance from transparency.
Openai was founded by Altman, Musk and Nine in 2015 as a non -profit research laboratory before Musk with Altman went out in an argument.
The head of Xai and Tesla is currently suing the start-up based in San Francisco and its boss for a company restructuring plan, which he claims to develop the profit for Openais Mission to “use humanity”.
The answers of the chatbots also help to disclose the restrictions of the AI models that they provide with electricity. The Chinese group Deepseeks Chatbot called its founder Lianfeng “an unconventional leader who prioritially prioritized creativity, passion and diverse perspectives against traditional experience”.
In contrast, US rivals corresponding to Claude, Lama and Google's Gemini didn’t know who was Wenfeng. This may be attributable to the incontrovertible fact that the businesses stopped behind these chatbots at the tip of last 12 months to gather training data before Deepseek jumped on global popularity in early 2025.
AI language models predict the subsequent probable word in a single sentence based in your training data. The latest AI chatbots also can browse on the Internet to win more sources. However, also they are depending on suitable English -language sources to attain answers. If information isn’t displayed in your training data, the model has little to work.
The chatbots that generated the answers to the Deepseek leader gave more general answers. At first, Meta Ai didn’t know who was Wenfeng, but once he said he was the CEO of Deepseek that he probably played a decisive role within the design of the corporate research and development of the corporate “.
This shows one other tendency of chatbots. Ki researchers said that chatbots have been trained to present plausible answers, to follow instructions and to generate things that folks need to hear.
Especially about their bosses.
