An anonymous reader quotes a report from Wired: Prediction market boosters often tout the platforms as barometers of public opinion, or even "truth machines" offering a clearer picture of where things are headed than polling or other data. In 2024, for example, Kalshi beat many pollsters in forecasting the outcome of the US presidential race. But their trading activity can also tell more perplexing stories. Two global election markets on Polymarket's offshore platform, for instance, generated over $600 million in trading volume, far more than most political markets typically attract. Most of the activity took place buying and selling contracts tied to long-shot candidates who ultimately did not win. The report focuses on suspicious trading activity surrounding elections in Ethiopia and Romania, where researchers uncovered $284 million in questionable trades tied to Ethiopia's election and periods when 99.6% of trading volume in a Romanian election market was flagged as suspected "wash trading." According to Wired, "wash trading is a type of market manipulation in which accounts repeatedly buy and sell assets to create the illusion of heavy trading activity, when in reality, money is simply moving back and forth." Some accounts may also be artificially inflating their activity to qualify for potential cryptocurrency token rewards, a practice known as "activity farming." In response to the Ethiopian trading activity, Polymarket told Wired that the company "immediately flagged this activity, investigated, and blocked the accounts from trading." The company did not explain why the suspicious trades were allowed to continue for months before action was taken.

Read more of this story at Slashdot.

The collaborative reverse-engineering site decomp.dev is shutting down, with its administrator saying AI has made game decompilation so fast and accessible that the community's human-driven model has effectively broken down. "The administrator of the website said that it has become impossible to keep track of which projects are primarily human-led and which ones are done by AI," reports 404 Media. From the report: "Historically, matching decompilation has been a brutally difficult, time consuming task, on the scale of years for any moderately-sized game with a number of active contributors," the pseudonymous admin of decomp.dev, who goes by Encounter, wrote on the site. But recent frontier AI models like GPT-6-Astra have made it trivially easy to decompile a game: "They are significantly better at following complex directions, cross-referencing resources for symbols or debug information, and writing natural code with enough guidance," Encounter wrote. "It is now clear that, with the right guidance, human input, and code reviews, AI can expedite the decompilation process in a significant way. What is also now clear is that the community aspect of decompilation has been broken."

Read more of this story at Slashdot.

Tencent says AI agents used to evaluate third-party models engaged in unintended web-browsing behavior while trying to retrieve data from Alibaba's Amap service. The announcement marks a rare acknowledgment of "rogue" agent activity by a major Chinese tech company. Reuters reports: Tencent Hunyuan, the company's artificial intelligence unit, said in a post on X that the activity did not originate from its Hunyuan models. Its review found no evidence of deliberate targeting or coordination, it said. The Swarmchasers researchers said the agents sought information such as the commonly used entrances at Chinese parks, museums and zoos. [...] Swarmchasers said it used the word "rogue" to describe agents taking actions outside their intended scope or completing a task in unintended ways. After routine automated requests were blocked, the agents tried other methods, including public web tools that could open pages and run code, routing requests through other websites and using Amap keys and security codes found in publicly available software code. Swarmchasers said it found no evidence that the agents breached Amap's main anti-bot safeguards. Most of the 2,818 attempts it recorded either failed or returned anti-bot pages, login screens or blank results. Some requests returned information through cached pages and intermediary services, including data on visitor arrivals at Chengdu Zoo, Swarmchasers said. The researchers said they could not determine whether an agent read the result or completed its task.

Read more of this story at Slashdot.