An anonymous reader quotes a report from TechCrunch: Fictional portrayals of artificial intelligence can have a real effect on AI models, according to Anthropic. Last year, the company said that during pre-release tests involving a fictional company, Claude Opus 4 would often try to blackmail engineers to avoid being replaced by another system. Anthropic later published research suggesting that models from other companies had similar issues with "agentic misalignment."

Apparently Anthropic has done more work around that behavior, claiming in a post on X, "We believe the original source of the behavior was internet text that portrays AI as evil and interested in self-preservation." The company went into more detail in a blog post stating that since Claude Haiku 4.5, Anthropic's models "never engage in blackmail [during testing], where previous models would sometimes do so up to 96% of the time."

What accounts for the difference? The company said it found that training on "documents about Claude's constitution and fictional stories about AIs behaving admirably improve alignment." Related, Anthropic said that it found training to be more effective when it includes "the principles underlying aligned behavior" and not just "demonstrations of aligned behavior alone." "Doing both together appears to be the most effective strategy," the company said.

A jury found Avraham "Avi" Eisenberg guilty on all three counts of fraud and manipulation in a $110 million crypto trade scheme using the Mango Markets platform. Axios: The case was the first known test for a jury to decide whether existing U.S. laws governing fraud and market manipulation apply to the world of decentralized finance (DeFi). The 28-year-old Eisenberg will be held to account for his actions on Oct. 11, 2022, when a series of trades he made intentionally boosted the price of Mango Markets' native token, MNGO, as well as the price of futures contracts.

He used the inflated futures holdings as collateral to borrow other cryptocurrencies on the platform, then quickly withdrew those assets and walked away from his collateral. Eisenberg never disputed the facts of the strategy but contended that what he did was legal and permitted by the DeFi protocol, a principle in the industry known as "code is law." U.S. laws apply to DeFi: "Avraham Eisenberg ran a con," prosecutors said Wednesday, during closing arguments, continuing its momentum from last week. The word "con" was used at least six more times in those remarks.