Unbiased AI-powered news
An inspection of 47 Home Office asylum cases from August to September 2025 found the majority likely incorrect and four in five made with insufficient evidence. GB News reported the findings from the Independent Chief Inspector of Borders and Immigration.
news.sky.comAn inspection by the Independent Chief Inspector of Borders and Immigration found the majority of 47 reviewed asylum decisions likely to be incorrect, GB News reported. Four in five of the cases spanning August to September 2025 were made with insufficient evidence.
The 44-page report titled An inspection of asylum casework from June-December 2025 stated that if the results were replicated across the system the quality of asylum decision-making was not in a good state.
A survey found 83.7 per cent of Home Office decision-makers believed senior managers prioritised quantity over quality. Individual decision-making units introduced targets and incentives that emphasised productivity, the report said. Quality has been the poor relation to productivity and quantity in operational priorities since at least 2020.
The review disclosed that 882 decision-makers left their roles from April 2024 to May 2025 after a recruitment push that added over 800 staff in early 2023. Candidates were not assessed against skills requirements for evidence analysis or decision-making, resulting in a 42.2 per cent attrition rate with employees leaving after an average of 11 months.
Artificial intelligence tools including Copilot and ChatGPT were used inappropriately in decision letters and interviews, the report found.
The Home Office stated that 4,000 cases show 94 per cent of decisions are considered correct against agreed criteria. Shadow Home Secretary Chris Philp said boat arrivals know their claim will almost certainly be approved regardless of its merits.
Single source — no framing comparison available.
thewire.inA coalition including Amnesty International and Save the Children called for governments to require safety checks on AI systems before release. The statement was issued one day before the United Nations holds its first global summit on AI governance.
airedale.futurecdn.netAlibaba directed employees to stop using Anthropic's Claude Code after the tool flagged connections from China. The company instructed staff to switch to its internal Qoder platform instead.
abcnews.go.comThe Trump administration removed limits on two Anthropic models last week that had been imposed the prior month. It separately asked OpenAI to delay a new series rollout.