Gedankenstrich.org

Das 2021 gegründete und im Mai 2026 mit 965 Milliarden US-Dollar bewertete KI-Unternehmen Anthropic (Claude) will an die Börse – und hat fast zeitgleich mit dieser Ankündigung einen Blog-Post veröffentlicht, der ein breites Medienecho hervorgerufen hat (z.B. Die ZEIT: »Anthropic fordert Pause bei Entwicklung von künstlicher Intelligenz«; tagesschau.de: »Anthropic fordert Pause bei KI-Entwicklung«).

In solchen und ähnlichen Fällen macht es Sinn, das Objekt der Berichterstattung im Original zu konsultieren – denn der Text »When AI builds itself: Our progress toward recursive self-improvement and its implications« liest sich doch deutlich differenzierter als es entlang der medialen Pointierung erscheint (und erweist sich in einigen Passagen schlicht als gut gemachtes Marketing-Material). Im letzten Drittel des Textes spannen die Autor:innen unter der Zwischenüberschrift »Possible Futures« drei Szenarien auf:

• »The trend stalls, but today’s AI capabilities are widely diffused. […]«
• »AI labs continue to see compounding efficiency gains. […]«
• »AI systems themselves become capable of full recursive self-improvement, and begin building their successors. If technical trends in advancing capabilities continue, and AI systems are able to develop the capabilities inherent to transformative human ingenuity, then it is plausible that AI systems could design and refine themselves. […]«

Primär für das dritte, auch im Anthropic-Text durchaus als voraussetzungsreich markierte Szenario diskutieren die Autor:innen schließlich geeignete gesellschaftliche Handlungsmöglichkeiten und heben dabei die Rolle des eigenen Unternehmens hervor:

»We believe it would be good for the world to have the option to slow or temporarily pause frontier AI development to enable societal structures and alignment research to keep up with the advance of the technology. The Anthropic Institute will conduct research—in collaboration with many others—and take actions to help build the systems that a credible slowdown or pause would require.

[…] A meaningful slowdown or pause would require multiple well-resourced labs at or near the frontier, in multiple countries, agreeing to stop under the same conditions. It would also require that each can verify that the others have actually stopped. […] None of this is necessarily impossible in principle—the world has built verification regimes for other complex technologies (e.g., the Intermediate-Range Nuclear Forces Treaty)—but those regimes took decades to build both the infrastructure and the trust. We don’t have that long. A unilateral pause by one lab, by contrast, is achievable immediately, but accomplishes much less: it would change who the front-runner is, but it would not create the wider deliberative process that is currently missing.

In the coming months, we will organize conversations where policymakers, researchers, civil society, and other AI companies can help answer some of the questions this piece raises, especially around full recursive self-improvement and how to create better options for coordination and deliberation. We’ll publish what comes out of it. The window to investigate the questions together is here, and people outside AI companies should be involved in this deliberation.«