What happened
The Metropolitan Police fired the head of its Special Constabulary, James Deller, following a misconduct hearing. Investigators used Palantir's Foundry AI system to detect that Deller had used his police-issued phone to copy intimate images sent via WhatsApp's 'view once' to his personal device in July and August 2025. The discovery occurred during 'Project Galatasaray,' a week-long operation aimed at identifying corruption and professional misconduct within the force. Deller has been added to the College of Policing's barred list, preventing him from serving in policing again.
The Metropolitan Police dismissed Special Chief Officer James Deller, the head of the force's volunteer officers, after a misconduct hearing concluded that he abused his position. The dismissal followed the discovery that Deller used his police-issued phone to copy two intimate images that had been sent to his personal phone using WhatsApp's 'view once' , which is designed to make images disappear after being opened. He then transferred these copied images to his personal device and deleted them from the work phone.
The misconduct was identified during 'Project Galatasaray,' a week-long operation that utilized Palantir's Foundry system, an artificial intelligence platform, to scan for corruption and professional misconduct within the Met. The force stated that investigators used lawful business monitoring combined with this AI technology to uncover the breach of trust. Cdr Pete Stevens, representing the Met, stated that Deller's actions were an abuse of the trust placed in him and that he is no longer able to serve in policing.
Deller has been added to the College of Policing's barred list, which means he is prohibited from serving in policing in the future. The Met emphasized that the vast majority of officers serve with integrity and that robust action is necessary against the small minority who undermine public trust, particularly within the Special Constabulary, where volunteers are expected to demonstrate the highest standards.
The use of AI in this investigation has sparked broader debate. In April, the Metropolitan Police Federation urged officers to be cautious with Met-issued devices off-duty after the force upgraded its Lawful Business Monitoring software. In May, Deputy Commissioner Matt Jukes told the London Assembly that officers were not under 'additional surveillance' but that the AI helped identify 'unacceptable, dishonest, corrosive behaviours.' However, Mopac, a procurement body, claimed the force failed to set out its procurement strategy, describing it as a serious breach of procedure.
Why it matters
This incident demonstrates the practical application of AI in internal law enforcement accountability, specifically using Palantir's Foundry platform to detect subtle misconduct that traditional monitoring might miss. It highlights the operational reality of AI-driven surveillance in public safety sectors, where algorithmic analysis of device usage can uncover violations of trust. The case also underscores the ongoing political and legal tensions surrounding Palantir's contracts with UK public bodies, including recent procurement disputes and pressure from political figures to review these partnerships. For the AI industry, it serves as a concrete example of how large-scale data integration and AI monitoring are being deployed in sensitive government contexts to enforce ethical standards.
This case is significant because it provides a verified, real-world example of AI being used to detect specific, subtle human misconduct within a major law enforcement agency. Unlike generic claims about AI efficiency, this report details how Palantir's Foundry system specifically identified the manipulation of 'view once' media files, demonstrating the technical capability of current AI monitoring tools in forensic and compliance contexts.
The incident highlights the intersection of AI deployment and public trust. While the Met argues the technology helps root out bad behavior, the controversy surrounding its procurement and the political pressure on the UK government to end contracts with Palantir suggest that the adoption of such tools is not without significant social and political friction. This friction is a critical factor for any organization considering similar AI-driven internal monitoring systems.
For the AI industry, this serves as a case study in the practical limits and applications of enterprise AI in the public sector. It shows that while AI can uncover violations, the surrounding legal, ethical, and political frameworks are equally important in determining the long-term viability and acceptance of such technologies in sensitive environments like policing.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').Why can ethical evaluation not be reduced to one model score?
What to watch next
Monitor the outcome of Palantir's lawsuit against Mopac regarding the procurement process, as this may set a precedent for how public bodies handle AI vendor contracts. Watch for further disclosures from Project Galatasaray regarding other potential misconduct cases identified by the AI system. Additionally, observe the political response in the UK to Palantir's continued involvement in public sector projects, particularly as pressure mounts from Labour MPs and trade unions to end government contracts with the company.
The legal battle between Palantir and Mopac is a key development to watch. Palantir is suing Mopac, alleging that the body unlawfully blocked a deal by considering the company's 'values and ethics' as an 'irrelevant consideration.' The outcome of this lawsuit could influence how public sector entities in the UK and potentially other jurisdictions approach the procurement of AI services from controversial tech firms.
Further results from Project Galatasaray may emerge, potentially revealing additional cases of misconduct identified by the AI system. If the operation leads to more dismissals or investigations, it could solidify the narrative that AI is an effective tool for internal policing, or conversely, fuel criticism if the findings are perceived as overreach or biased.
Political pressure on the UK government regarding Palantir contracts is likely to intensify. With Prime Minister Andy Burnham facing calls from Labour MPs and trade unions to end government contracts with Palantir, any policy shifts or public statements from the government could impact Palantir's operations in the UK and serve as a bellwether for AI procurement policies in other countries.