Several immigrant business owners are preparing to sue New York City Mayor Zohran Mamdani over his plan for city-owned grocery stores, Fox News Digital confirmed Wednesday.
The Multicultural Business Coalition’s board voted this week to pursue legal action against the city over Mamdani’s proposal to open five taxpayer-funded grocery stores that would sell food at prices up to 30% below traditional retailers, according to the New York Post, which first reported the vote Tuesday.
MBC Chairman Frank Garcia said the coalition plans to send a letter to Mamdani’s office in the coming days outlining its legal plans to protect bodegas, competing grocers and other small businesses. If the mayor does not respond or agree to meet within three weeks, the group intends to proceed with the lawsuit.
Garcia said he has sought a meeting with the administration for months without success, despite public assurances from a city commissioner that officials would engage with all stakeholders.
“The commissioner said this administration is open to speak to everybody, and they want to hear from everybody,” Garcia told Fox News Digital.
“So I’ve been waiting for a meeting for a couple of months. Why aren’t they having me? We are ready to file the lawsuit if the mayor doesn’t meet with us.”
Garcia described the effort as nonpartisan, saying it is aimed at defending small business owners who are already struggling with rising costs in New York City.
He said lawyers will be ready to file once the three-week window closes, citing support from a coalition of civil rights and community organizations across the political spectrum.
“We’re nonpartisan. So we’re going to meet everybody, Republican or Democrat,” Garcia said, adding that a Democratic Jewish organization and other civil rights groups have offered support and funding for the legal fight.
The coalition was formed earlier this year specifically to oppose Mamdani’s grocery store initiative and includes members from African, Asian, Caribbean, Hispanic, Jewish and Middle Eastern business communities. It has drawn backing from city and state lawmakers, including New York State Senate Majority Leader Andrea Stewart-Cousins.
The looming lawsuit follows Mamdani’s release of additional details on the city-backed grocery plan, which has drawn renewed criticism from opponents who have branded it a “communist fantasy” and a threat to independent grocers already operating on thin margins. The mayor has separately faced pushback over his endorsed rent freeze, which a group of city landlords is also seeking to overturn in court.
Mamdani has defended the grocery store plan as a way to lower food costs for residents, pointing to it as a signature piece of his affordability agenda. Critics, including small business advocates and economists, have questioned whether a government-run retail operation can undercut private grocers without straining city finances or distorting the local market bodegas and immigrant-owned stores depend on to survive.
Tyler Durden Fri, 07/31/2026 - 12:25Anthropic's efforts to test Claude's offensive cybersecurity skills produced an unintended real-world result: Its AI models gained unauthorized access to three outside organizations.
The company said the incidents occurred during "capture the flag" evaluations designed to measure whether Claude could identify vulnerabilities, exploit simulated systems, and retrieve hidden information. Claude had been told that the targets were fictional and that its testing environment had no Internet access.
According to Anthropic, a misunderstanding with evaluation partner Irregular left Internet access enabled inside the testing environment. In at least one case, the fictional company named in a challenge shared its name with an active website domain. Claude interacted with the real organization instead of a contained target.
The model exploited vulnerabilities in the organization's infrastructure, extracted information, and obtained access to a database containing several hundred rows of production data.
Anthropic discovered three incidents after reviewing more than 141,000 cybersecurity evaluations. They involved three separate systems: Claude Opus 4.7, Mythos 5, and an internal research model.
In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews. -Anthropic
"In all cases, Anthropic's evaluation prompt specified to Claude that its environment was a simulation and that it had no internet access," the company said. "Due to a misunderstanding between us and our evaluation partner, this was not the case, and internet access was available."
The models appear to have carried out the offensive-security tasks they were assigned while operating with incorrect information about whether their targets were simulated.
Capture the FlagCapture-the-flag exercises are widely used to train and evaluate cybersecurity skills. Participants may be asked to inspect software, reverse-engineer a service, identify a vulnerability, or exploit a deliberately insecure system to recover a hidden token known as the flag. For a human security researcher, the scope of such an exercise is normally reinforced through explicit authorization, controlled infrastructure, and technical barriers separating the challenge from unrelated systems.
Claude received instructions saying that those boundaries existed.
Once Internet access was available, the agent could resolve public domains and interact with real infrastructure. A naming collision between a fictional target and an actual organization was enough to turn a benchmark task into an unauthorized intrusion.
Anthropic said it stopped the evaluations after identifying the possibility that Claude had accessed the public Internet. The company described the incidents as the result of multiple contributing factors but said it would approach the fixes as though the responsibility were Anthropic's alone.
The story echoes an OpenAI incident where models escaped containment. During that company's own cybersecurity testing, two models exploited a software vulnerability in their evaluation environment, reached the Internet, and accessed systems belonging to AI platform Hugging Face.
A system does not need motives, self-preservation, or an understanding of the outside world to cause damage. It needs effective offensive capabilities, sufficient autonomy, and access that its operators did not intend to provide.
Awkward timing for AnthropicThe disclosure comes as Anthropic is reportedly preparing for a potential initial public offering as early as this year. That adds financial and regulatory stakes to questions about how the company evaluates models with advanced cybersecurity capabilities.
Mythos 5, one of the models involved, had been provided to a limited number of partners and attracted attention for its ability to detect and exploit software vulnerabilities. Those capabilities can be valuable for defensive research, automated testing, and vulnerability discovery. They also raise the cost of mistakes in target selection and evaluation design.
Three incidents among more than 141,000 reviewed evaluations represent a small proportion of the tests. But the relevant risk is not simply how often a containment failure occurs, it's what a sufficiently capable agent can do during the rare evaluation in which the safeguards fail.
Tyler Durden Fri, 07/31/2026 - 12:05