OpenAI против Anthropic: почему гонка ИИ тревожит исследователей и Цукерберга

Wired published an analytical piece on the intensifying competition in the AI industry, gathering three key concerns: scientists warn that OpenAI and Anthropic are moving too fast, Mark Zuckerberg raised the question of who will control AI, and Berlin-based startup Black Forest Labs announced its entry into robotics.

Why researchers fear the pace of the race

The main grievance of the scientific community — the two leading AI laboratories are scaling up capabilities faster than understanding of their systems' safety is developing. Competitive pressure between OpenAI and Anthropic, critics argue, forces companies to sacrifice caution for the speed of releasing new products and models. Concern is growing as each successive release becomes more powerful than the previous one, while unified industry safety evaluation standards have still not been developed. Particular concern is raised by the absence of public independent audits: neither company discloses full audit results before a system is released.

As Wired reports, the research community is voicing concern that actual safety verification in practice is being shifted onto real users — after products are released, not before. OpenAI and Anthropic set the pace not only for themselves but for the entire industry: smaller players are forced to follow the leaders to avoid losing competitiveness. It is telling that criticism increasingly comes not only from academic circles, but also from specialists working inside the industry itself.

Who will own artificial intelligence?

Mark Zuckerberg, CEO of Meta, raised a question that goes beyond simple safety concerns: who will ultimately really control the planet's key AI infrastructure. Meta consistently bets on open models in the Llama series as a systemic counterweight to the closed developments of OpenAI and Google. By Zuckerberg's logic, public availability of model weights reduces the risk of monopolistic control over the technology.

"Open AI is the best outcome for the world" — this position of

Mark Zuckerberg, CEO of Meta, became the leitmotif of his public speeches about the future of the AI industry; Wired considers it a key alternative to the closed approach of OpenAI and Anthropic.

Critics, however, object: openness of weights also lowers the threshold for misuse. In the context of the race between the two leading laboratories, Meta's position is not only a declaration of principles, but also a competitive move: open AI strengthens the platform's ecosystem without the need to build its own closed monopoly. According to Wired, concentration of control over AI models is increasingly becoming a political question — discussed by both CTOs and regulators.

Black Forest Labs: from image generation to robots

Berlin-based startup Black Forest Labs, which created the FLUX family of generative models for working with images, is expanding its activities toward robotics — as reported by Wired. Black Forest Labs became one of the key European suppliers of open models for image generation. For a company that built its reputation in the niche of visual AI, entering physical robotics represents a significant strategic shift.

The transition of generative startups into the physical world is becoming a sustained trend: multimodal models are increasingly viewed as the "intelligence" for robotic systems. Black Forest Labs is entering a space where Google DeepMind's physical AI divisions, OpenAI, and a number of industrial partners already compete. For the Berlin startup, this is a bet on expansion beyond the image generation market — into territory where software models intersect with the real physical world.

What this means

The competition between OpenAI and Anthropic is no longer an internal matter of two laboratories — it is shaping the regulatory agenda, provoking Meta to redefine the rules through openness, and drawing new participants like Black Forest Labs into the race. The question of the pace of development and who ultimately owns artificial intelligence is moving from the technical to the political, and will remain central to the industry for years to come.

*Meta has been recognized as an extremist organization and is banned in Russia.