After the Mythos ban, acceleration AI ethics proposes a third way: Safety through innovation

The recent suspension of Anthropic’s Mythos and Fable models brings the conflict between AI innovation and safety into public view: some favor innovation prohibitions while others are skeptical of safety constraints. A new paper in Law, Ethics and Technology by philosopher James Brusseau resolves the dilemma with “acceleration AI ethics,” a framework for answering AI risks with further innovation. Using the Telus GenAI customer support agent as a case study, the article argues that responsible AI does not have to choose between moving fast and advancing safely. Instead, the safety risks generated by innovation are overcome by the same innovative forces that created them.

The U.S. government directive restricting access to Anthropic’s Mythos 5 and Fable 5 models has turned a technical question into a public one: when a powerful AI system appears risky, should society stop the technology, or stand aside while the AI rushes ahead?

A new article by James Brusseau, published in Law, Ethics and Technology, argues that the answer does not have to be a simple choice between defensive precaution and reckless advancing. The paper defines “acceleration AI ethics” as a third approach: safety is maximized through innovation. In responsible AI, the risks created by innovation are met by new technical, organizational, and ethical innovations that move forward to manage those risks.

“The Mythos episode is the kind of dilemma this paper addresses,” says Brusseau, who is affiliated with Pace University in New York City and the University of Trento in Italy. “The standard debate asks us to choose between innovation and safety. Acceleration ethics tries to dissolve their opposition. It asks how the underlying ethical forces driving innovation can subsequently produce safety.”

The article identifies five elements of acceleration AI ethics. First, innovation solves innovation problems: risks created by AI should be answered with further AI and engineering advances. Second, innovation is intrinsically valuable: technical creativity has worth before its downstream uses are known. In this way, AI engineering resembles artistic creation. Third, and most significantly, the unknown is understood as innately magnetic and encouraging: uncertainty is not a warning, but a reason to explore. Fourth, governance is decentralized: rules and safety practices emerge from actual use and users, rather than only from pre-emptive external control. Fifth, ethics is embedded: ethicists and engineers work together from the beginning, converting human concerns into design challenges.

To show how the framework works in practice, the paper examines Telus’s GenAI customer support tool. The tool was built to answer customer questions across the telecommunications company’s services, and its development raised familiar generative AI concerns including hallucinations, prompt hacking, and privacy. Within the ethics of acceleration, these risks became innovation incentives: they served both as reasons and as ways for more technological advance. Concretely, a second agent was developed out of the first as an automated adversarial testing system: AI was used to challenge AI, probing the customer-support agent for vulnerabilities so that risky outputs could be flagged for human review and technical improvement. The answer, in other words, to the initial innovation risk was a subsequent round of innovation as propelled and directed by the risks themselves.

Ultimately, the Mythos ban shows why the dilemma between innovation and safety is urgently important. Brusseau’s paper reframes that dilemma by showing how safety concerns can become directions for accelerating responsible innovation.

This paper, “Acceleration AI ethics and the Telus GenAI conversational agent,” was published in Law, Ethics and Technology.

Brusseau J. Acceleration AI ethics and the Telus GenAI conversational agent. Law Ethics Technol. 2026(2):0006, https://doi.org/10.55092/let20260006..

Author: James Brusseau

Author affiliations: Philosophy Department, Pace University, New York, USA; Department of Information Engineering and Computer Science, University of Trento, Trento, Italy

Keywords: generative AI; case study; innovation; safety; acceleration ethics; Telus

Media contact: jbrusseau@pace.edu