How far can MI5 agents go in the field? This is the question being asked by four UK NGOs in an ongoing legal challenge against MI5’s policy authorising agents or informants to participate in crime while undercover. The concern is not only that the UK’s domestic intelligence agency is caught up in human rights abuses, but that the government has actively avoided legislating further on this issue despite urgent calls to do so.


In seeking to keep the country safe, just how far can the intelligence agencies go? Can they get away with murder? This is the question being asked in an ongoing legal challenge brought by four UK NGOs:  Reprieve, Privacy International, the Committee on the Administration of Justice, and the Pat Finucane Centre. The challenge is to a policy run by the Security Service — the UK’s domestic intelligence agency, known as MI5 — to authorise agents or informants to participate in crime while undercover. In this article I’ll explain what is at stake, and why this is so important for human rights and the rule of law in the UK.

The Government has sought to keep this policy secret for almost 30 years. It only came to light in a separate case, brought by Privacy International, in which the Government revealed that the UK’s chief intelligence watchdog had been overseeing an unknown area of intelligence activity since 2012. The guidance has since been made public through disclosure, but the most important sections — on how far agents can go and what crimes they can be authorised to commit — remain secret.

Late last year, the Investigatory Powers Tribunal — the UK’s first-instance court for claims against the intelligence agencies — found against us, refusing requests to declassify more of the policy. We have permission to appeal. But for the first time in the Tribunal’s 20-year history, two of its five-judge panel issued dissents, in notably strong terms, describing the Government’s attempts to demonstrate a legal basis for the policy as “extraordinary” and “fanciful”.


Is MI5 becoming mixed-up in rights abuses?

The ability to run agents is vital to the work of any intelligence service. Recruiting agents within proscribed groups helps gather intelligence and thwart serious crimes. But how far can the agencies go? In seeking to stop crime, can agents commit crimes themselves? And might MI5 officers become complicit in human rights abuses committed by the agents they run?

These questions have serious implications. During the conflict in Northern Ireland, the UK Government ran agents in both Nationalist and Loyalist groups to gain intelligence and thwart attacks. This risked UK forces becoming mixed up in human rights abuses as well however, and reviews eventually concluded that the unrestricted use of agents ultimately made the conflict worse, not better.

In 1989, Belfast lawyer Pat Finucane was shot fourteen times as he sat down to dinner with his family. It emerged that the Loyalist group responsible had been infiltrated by British intelligence. Following a review of the killing in 2012, then-Prime Minister David Cameron admitted there had been “shocking levels of state collusion” in Mr. Finucane’s murder. The review also found that the Ministry of Defence exaggerated the usefulness of their agents within the group that killed Mr. Finucane, that claims made to justify their actions were “utterly wrong”, and that their activities “had not generally saved lives”.


How far can agents go?

As it states, the policy provides MI5’s “own procedure for authorising the use of agents participating in crime” by which the agency seeks to “secure and maintain access to intelligence”. The Government is fighting to keep the limits of the policy secret – in other words, seeking to keep hidden whether agents working for MI5 can commit serious crimes, even murder or torture, in order to maintain their cover.

Former officials have been less reticent in describing the limits of the policy in public. During a BBC Radio 4 Today Programme interview last year, the former Director-General of MI5 Lord Evans was asked whether the policy might permit “punishment beatings”. He said: “It’s not possible for that to happen without…[pause]…there are no specific rules on exactly which crimes”. So the policy may have no limits at all, leaving MI5 officers potentially complicit in abuses committed by the agents they run.

This raises alarming questions from the earliest days of the ‘war on terror’. In 2018, Parliament’s Intelligence and Security Committee identified hundreds of cases of UK involvement in abuses committed by foreign intelligence partners from 2001 onwards, finding that the agencies engaged in “simple outsourcing of action they knew they were not allowed to undertake themselves”. Might this policy be allowing MI5 to do the same at home, or perhaps worse? The Government claimed in an argument before the Investigatory Powers Tribunal that the state, in tasking an agent, “is not the instigator of that activity and cannot be treated as responsible for it” — so the answer may be ‘yes’.

The UK Government claims that to publish further details would make the job of recruiting agents more difficult. Yet what the Government allows its agents to do is limited by UK law. As the High Court found in respect of the ongoing ‘Spy Cops’ scandal — in which undercover police officers were found to have infiltrated protest groups and in some cases allegedly committed sexual assaults against members — “there can be no license for torture or for any other inhuman or degrading treatment”.


Does the policy have any legal basis?

Whatever the policy actually says, it has no basis in law. The UK’s overseas intelligence agency — the Secret Intelligence Service, known as MI6 — has a statutory regime to effectively authorise the commission of offences abroad: section 7 of the Intelligence Service Act 1994. There’s no such regime for MI5 at home though, and no powers of this kind have ever been approved by Parliament.

In fact, UK Government memos quoted in the 2012 report on Pat Finucane’s murder make clear that when this policy was drawn up, senior officials believed that new legislation, not guidance, was urgently needed. As one memo noted however, the Government decided against asking lawmakers for new legal powers on the basis of “grave reservations about opening up such a sensitive area to Parliament when the slenderness of the Government’s majority could not guarantee a satisfactory outcome”.

When in 2000 a different Government introduced an authorisation process for the use of covert agents with the Regulation of Investigatory Powers Act, it did not include a process for authorising participation in crime.  So it remains the case that MI5 tells its agents to commit crimes — potentially involving life and death — without any authorisation by Parliament. This is deeply unacceptable in a democracy whose intelligence agencies are supposed to act under the sanction of Parliament and within the rule of law.

The Government claims that section 1 of the 1989 Security Service Act provides the authority for its policy of sanctioned law-breaking. However, this provision just states MI5’s function, i.e. the protection of national security, as it placed the agency on a statutory footing for the first time in its history.

And if the Government can use this to establish a regime of authorised law-breaking by the intelligence agencies, what else might it do on the same legal basis? The two judges who found in our favour at the Investigatory Powers Tribunal agreed. One described the Government’s legal justification as setting “a dangerous precedent”, while another asked of the Government’s claimed legal authority, “can this possibly be correct? Where does it end? What other powers does MI5 have as a result of [section 1 of the 1989 Act]?”

In making these arguments, the Government is swimming against the tide of intelligence agency oversight across Europe. The European Convention on Human Rights requires that the agencies operate within clear legal limits and that their powers (if not specific operations) are clearly signposted to the public and certainly set out in statute. We’ve moved on from when MI5, as described by former officer Peter Wright in his 1987 memoir Spycatcher, “bugged and burgled… while pompous bowler-hatted civil servants in Whitehall pretended to look the other way”. For these highly intrusive powers to authorise crimes by MI5 agents however, the UK Government seems to think that nothing has changed.


The need to safeguard against abuses

The Government also claims that, whatever the policy says, it has been overseen by the Investigatory Powers Commissioner since 2012 and that he has given MI5 a clean bill of health. Yet we know that the Prime Minister told the Commissioner not to review the legality of the policy or whether cases should be referred to the independent Crown Prosecution Service for criminal proceedings. Meanwhile most of the Commissioner’s findings about the operation of the policy remain classified. We do know that he recently expressed concern that “MI5 lack reliable central records around [participation in crime] activity and that there is no consistent review process” — raising questions about the effectiveness of this oversight.

During the early years of the ‘war on terror’, as the agencies became increasingly involved in torture and rendition around the world, MI6 officers reportedly expressed an “uneasy feeling of operating in a legal wilderness”. Secret guidance to staff permitted abuses to take place, and years of revelations have made clear the damage done by operating without clear safeguards against abuses. Without real limits on what MI5 agents may do when breaking the law, we risk making the same mistakes again.

Involvement in serious abuses of human rights damages the agencies’ ability to do their vital work of keeping the UK safe – something that the agencies themselves reportedly recognised as evidence emerged of their involvement in torture and rendition overseas. MI5 should be using the right tools to identify and eliminate threats to the UK. But these tools need to be lawful as well as proportionate and accountable. For MI5’s use of agents who participate in crime, we don’t yet have that guarantee.

Privacy Preference Center