Vested.

We need to stop AI developing without human input, says Anthropic co-founder

5h ago · UK · primary source: feeds.bbci.co.uk

Multi-source synthesis by Vested from 2 sources. Every numeric and quoted claim traces to a cited source body (see methodology).

Anthropic's widely used chatbot Claude already runs on code that is 80 percent self-written, with the potential to reach full autonomy within two years, the company's co-founder has stated[1][2]. The AI firm is simultaneously preparing for a public stock market debut with an estimated valuation approaching $1 trillion[1][2].

Co-founder Jack Clark told the BBC that achieving 100 percent self-written code for Claude is possible within two years and "would have huge implications"[1][2]. He warned the AI industry currently lacks a regulatory "brake pedal" to match its accelerating development pace, drawing a parallel to the historical need to regulate the oil industry[1]. "You want the option to be able to take your foot off the gas and put your foot on the brake", Clark said. "Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal"[1].

Despite these concerns, Anthropic recently welcomed a relatively hands-off AI executive order from US President Donald Trump, which did not mandate government safety testing[1]. Major AI companies, including Anthropic, OpenAI, and Google, have not announced any pauses to their research efforts[1].

Anthropic's leadership, including CEO Dario Amodei and Clark, has been outspoken about potential risks from AI advancements[2]. The company's rapid growth since its founding five years ago has positioned it for what could be one of the most valuable stock listings in history[1][2]. Clark stated the motivation for publicly discussing AI capabilities is to inform the world about developments inside these companies, not merely to burnish Anthropic's reputation[1].

Context we found (10)

  • en.wikipedia.org — https://en.wikipedia.org/wiki/Artificial_general_intelligence ↗
    Artificial general intelligence (AGI) is a hypothetical type of artificial intelligence that matches or surpasses human capabilities across virtually all cognitive tasks. Beyond AGI, artificial superintelligence (ASI) would outperform the best human abilities across every domain …
  • en.wikipedia.org — https://en.wikipedia.org/wiki/AI_agent ↗
    In the context of generative artificial intelligence, AI agents (also referred to as compound AI systems or agentic AI) are a class of intelligent agents that can pursue goals, use tools, and take actions with varying degrees of autonomy. In practice, they usually operate within …
  • en.wikipedia.org — https://en.wikipedia.org/wiki/Grok_(chatbot) ↗
    Grok is a generative artificial intelligence chatbot developed by SpaceXAI. It was launched in November 2023 by Elon Musk as an initiative based on the large language model (LLM) of the same name. Grok has apps for iOS and Android and is integrated with the X social network and T…
  • en.wikipedia.org — https://en.wikipedia.org/wiki/Google_DeepMind ↗
    Google DeepMind, trading as Google DeepMind or simply DeepMind, is a British-American artificial intelligence (AI) research laboratory which serves as a subsidiary of Alphabet Inc. Founded in the UK in 2010, it was acquired by Google in 2014 and merged with Google AI's Google Bra…
  • arxiv.org — https://arxiv.org/abs/2604.23178v1 ↗
    LLM-as-a-Judge has become the dominant paradigm for evaluating language model outputs, yet LLM judges exhibit systematic biases that compromise evaluation reliability. We present a comprehensive empirical study comparing nine debiasing strategies across five judge models from fou…
  • arxiv.org — https://arxiv.org/abs/2604.12076v1 ↗
    The Identifiable Victim Effect (IVE) $-$ the tendency to allocate greater resources to a specific, narratively described victim than to a statistically characterized group facing equivalent hardship $-$ is one of the most robust findings in moral psychology and behavioural econom…
  • arxiv.org — https://arxiv.org/abs/2512.24526v1 ↗
    This paper investigates how Large Language Models (LLMs) from leading providers (OpenAI, Google, Anthropic, DeepSeek, and xAI) can be applied to quantitative sector-based portfolio construction. We use LLMs to identify investable universes of stocks within S&P 500 sector indices …
  • arxiv.org — https://arxiv.org/abs/2509.11295v2 ↗
    Developing effective prompts demands significant cognitive investment to generate reliable, high-quality responses from Large Language Models (LLMs). By deploying case-specific prompt engineering techniques that streamline frequently performed life sciences workflows, researchers…
  • arxiv.org — https://arxiv.org/abs/2505.03553v1 ↗
    Inconsistent outputs and hallucinations from large language models (LLMs) are major obstacles to reliable AI systems. When different proprietary reasoning models (RMs), such as those by OpenAI, Google, Anthropic, DeepSeek, and xAI, are given the same complex request, they often p…
  • en.wikipedia.org — https://en.wikipedia.org/wiki/Anthropic ↗
    Anthropic PBC is an American artificial intelligence (AI) company headquartered in San Francisco, California. It has developed a series of large language models (LLMs) named Claude and has a focus on AI safety. Anthropic was founded in 2021 by former members of OpenAI, including …

Sources cited (2)

  1. bbc.com B · newspaper — https://www.bbc.com/news/articles/cx2124z7g45o?at_medium=RSS&at_campaign=rss ↗
  2. bbc.com B · newspaper — https://www.bbc.com/news/articles/cx2124z7g45o?at_medium=RSS&at_campaign=rss ↗
Spot something wrong? Report an issue