MI5 has warned that a Chinese research institute is helping China's intelligence service exploit British academic research. The more consequential story may be what Britain's own National Security Act now allows the state to do in response.
There are two stories unfolding in Britain at the same time.
The first is the intelligence story.
On September 30, MI5 issued what it described as a Security Service Espionage Alert concerning the China General Technology Research Institute, or CGTRI. MI5 says the institute has “very strong ties” to China's Ministry of State Security and that its primary purpose is to fund research that directly improves the MSS's technical capabilities for espionage.
More than 100 UK-linked academics, MI5 says, have contributed to CGTRI-funded research involving subjects including artificial intelligence, cybersecurity, covert communications and steganography. MI5 also warned that some academics may not have known that CGTRI was ultimately involved in funding the projects.
The second story is about Britain itself.
Because Parliament has spent the past several years watching the Chinese state expand its national-security laws — particularly in Hong Kong — Britain has frequently presented itself as the country defending liberal democratic values against excessive state power.
But Britain has also been building its own national-security legal architecture.
The National Security Act 2023, significantly expanded by the National Security (State Threats) Act 2026, gives the British government an unusually broad collection of offences and powers aimed at foreign intelligence services and bodies involved in foreign-power threat activity.
And the new legislation now reaches beyond espionage in the traditional sense.
It reaches into financial benefits, information, assistance, organisational support and, in certain circumstances, speech supportive of a designated body.
That does not make British national-security law identical to Hong Kong's.
It does, however, make the British government's criticism of expansive national-security legislation considerably more complicated.
Start with what MI5 actually said
The CGTRI warning is serious.
MI5 says the institute funds academic research that directly improves the technical capability of China's Ministry of State Security.
It specifically identifies AI and cybersecurity among the research areas involved and says more than 100 UK-linked academics have contributed to research projects funded through CGTRI.
MI5's advice to universities is blunt: review existing and planned collaborations with CGTRI and establish the ultimate source of funding when working with Chinese institutions.
The agency also says academics should take independent legal advice about continuing research ultimately funded by CGTRI.
This is important because the warning is not simply saying:
China is interested in British research.
That is hardly surprising.
The accusation is considerably more specific: MI5 says CGTRI is functioning as a channel through which Chinese intelligence obtains technological capabilities relevant to espionage.
The Chinese government has rejected such espionage accusations. Reuters reports that Beijing's embassy in London dismissed the claims as unfounded and urged Britain not to damage academic cooperation.
And MI5 has not published the intelligence underlying its assessment.
That is normal for an intelligence service. Intelligence agencies routinely protect sources and methods.
But it creates an important distinction between an intelligence assessment and a court-tested factual finding.
MI5 says this is what its intelligence shows.
A criminal court, by contrast, applies evidential and procedural rules to determine whether the elements of a particular offence have been proved.
That distinction becomes crucial when examining the legislation.
Section 17 is broader — but not quite as broad as the viral example suggests
One of the most striking provisions of the original National Security Act is section 17.
It creates offences concerning obtaining material benefits from a foreign intelligence service.
Under section 17(1), a person commits an offence if they obtain, accept or retain a material benefit provided by or on behalf of a foreign intelligence service, while knowing — or, having regard to other matters known to them, ought reasonably to know — that it came from such a service.
And section 17(2) goes further.
A person can commit an offence simply by agreeing to accept a material benefit that is to be provided by or on behalf of a foreign intelligence service, provided the requisite knowledge condition is satisfied.
The maximum penalties are significant:
- up to 14 years for an offence under section 17(1);
- up to 10 years for an offence under section 17(2).
The legislation also defines material benefits broadly.
They may include financial benefits, things with the potential to result in a financial benefit, and information.
The government's own explanatory notes make clear that the provision was deliberately designed to cover benefits beyond straightforward cash payments.
They give examples involving payments, gifts with monetary value, beneficial contracts and information.
That is a powerful provision.
But there is an important qualification.
The coffee example is too broad
The claim that accepting an ordinary coffee from someone connected to a Chinese research institute would automatically amount to a section 17 offence goes further than the statute supports.
Section 17 requires the benefit to be provided by or on behalf of a foreign intelligence service.
A Chinese research institute merely being accused by MI5 of having links to an intelligence service does not automatically mean that every coffee, lunch or conversation involving one of its employees is a benefit provided on behalf of that intelligence service.
There are also exclusions for certain benefits provided as reasonable consideration for lawful goods or services.
So the important point is not that British academics now have to fear accepting every cup of coffee offered by a foreign researcher.
The more significant point is this:
British law has expressly criminalised the acceptance — and even agreement to accept — certain material benefits from foreign intelligence services where the required knowledge test is satisfied.
And Parliament has defined "material benefit" broadly enough to include information.
That is considerably more expansive than the traditional image of espionage as stealing secrets.
Then came the 2026 expansion
This is where the story becomes much more interesting.
On July 8, 2026, new provisions inserted sections 17A, 17B and 17C into the National Security Act.
These provisions create a new legal category around designated bodies involved in foreign-power threat activity.
Under section 33A, the Secretary of State can designate a body if the Secretary of State:
- reasonably believes the body is or has been involved in foreign-power threat activity; and
- considers designation necessary to protect the UK's safety or interests.
There is parliamentary oversight.
Regulations designating a body normally require approval by both Houses of Parliament.
There is also an urgent procedure under which designation can temporarily take effect without prior approval, although it expires after 40 days unless both Houses approve it.
So the claim that the Home Secretary could simply declare an organisation dangerous and Parliament has absolutely no role is not accurate.
Parliament does have a role.
But the crucial question is what happens after designation.
Britain has now criminalised certain forms of support for designated bodies
Section 17A is extraordinary in its wording.
It says a person commits an offence if they:
invite support for, or express an opinion or belief that is supportive of, a designated body
and do so for a prohibited purpose.
The "prohibited purpose" is a purpose that the person knows, or — considering other matters known to them — ought reasonably to know is prejudicial to the safety or interests of the United Kingdom.
The maximum sentence is 14 years' imprisonment.
That is not a trivial provision.
It means that Parliament has now created a national-security offence in which expressing an opinion or belief supportive of a designated body can, in circumstances specified by the statute, become criminal conduct.
That is a significant departure from the conventional espionage model.
Traditional espionage law is generally concerned with what someone does:
- obtaining classified information;
- passing secrets;
- assisting an intelligence service;
- conducting covert activity;
- sabotaging infrastructure.
Section 17A introduces something different.
Speech itself can become part of the offence.
But again, precision matters.
It does not mean that every positive comment about an organisation automatically becomes criminal.
There must be a designated body and the statutory "prohibited purpose" must also be established.
The offence therefore has legal conditions beyond merely saying something nice about an organisation.
Imagine CGTRI were designated
This is where the hypothetical becomes genuinely interesting.
The present MI5 warning does not designate CGTRI under section 33A.
It is an intelligence warning.
But imagine that a future government formally designated CGTRI under the new legislation.
At that point, section 17A could become relevant.
Suppose an academic publicly argued:
"CGTRI has made a positive contribution to British research."
That statement by itself would not automatically establish a criminal offence.
The prosecution would still have to establish the statutory requirements, including the prohibited-purpose element.
But the law's wording is undeniably broad enough to encompass an "opinion or belief that is supportive" of a designated body.
That is the part that deserves serious public scrutiny.
Because national-security legislation has crossed an important conceptual line when an individual's expression of support can become criminal conduct under specified circumstances.
And there is another new offence: assisting designated bodies
Section 17B goes beyond speech.
It criminalises conduct intended to materially assist a designated body in carrying out UK-related activities.
It also covers conduct that is likely to materially assist such a body where the person knows — or ought reasonably to know — that it is likely to do so.
The legislation specifically says potentially assisting conduct can include providing:
- information;
- goods;
- services;
- financial benefits.
And it can happen directly or indirectly.
The architecture is therefore becoming broader:
espionage → assistance → material benefits → designated organisations → support and speech.
That is why the 2026 amendments deserve attention independently of the China controversy.
But does designation mean the accusation is never tested in court?
This is another area where the viral argument needs refinement.
It is true that designation itself is based on the Secretary of State's reasonable belief, rather than a criminal conviction establishing beyond reasonable doubt that the body committed foreign-power threat activity.
The designation process is an executive decision subject to parliamentary approval through the statutory instrument process.
But it is too simplistic to say that once designated, "the accusation never has to be proven in your trial."
A criminal prosecution under section 17A would still have to establish the elements of the particular offence.
The court is not simply convicting someone because a minister once said an organisation was dangerous.
The designation establishes the legal status of the organisation for purposes of the statutory regime.
The prosecution still has to prove the defendant's conduct and the relevant mental or knowledge requirements.
That distinction matters enormously.
Yet the designation system creates a real due-process question
The deeper issue is not whether Britain has evidence.
MI5 presumably has intelligence behind its assessments.
The issue is where that evidence is tested, and by whom.
National-security systems inevitably involve intelligence that cannot be publicly disclosed.
The government says this is necessary to protect intelligence sources and methods.
The counterargument is equally familiar: if secret intelligence becomes the foundation for severe restrictions on individuals or organisations, democratic oversight becomes more difficult because the affected person may not be able to see or challenge the underlying material.
That is not uniquely British.
It is a recurring problem in national-security law around the world.
But Britain's new framework makes the question increasingly relevant because designation can activate offences carrying extremely serious penalties.
The Hong Kong comparison needs more nuance
This is where the comparison with Hong Kong becomes politically charged.
Britain has spent years criticising Hong Kong's National Security Law.
The United Kingdom has described China's actions in Hong Kong as damaging to rights, freedoms and the rule of law.
Meanwhile, Hong Kong's national-security regime has produced prosecutions involving foreign-force offences.
Jimmy Lai is the most prominent example.
Lai was convicted of two counts of conspiracy to collude with foreign forces and one count of publishing seditious material. In February 2026, he was sentenced to 20 years in prison.
The prosecution's case concerned alleged cooperation with foreign actors and efforts to influence governments.
That system has been heavily criticised internationally.
But the comparison should not be reduced to:
Hong Kong requires evidence; Britain requires none.
That is not a fair description of either legal system.
Hong Kong courts conduct criminal trials and convictions require the prosecution to establish the charged offences according to the applicable legal standard.
Britain does the same.
The difference worth examining is elsewhere.
Britain's new law creates an administrative designation mechanism which can determine that a body falls within a special national-security regime before criminal liability is subsequently assessed in individual cases.
That is a significant concentration of executive power, even though parliamentary approval and legal challenges exist.
The real danger is not that Britain suddenly became Hong Kong
That would be too simplistic.
Britain still has a different constitutional system, independent courts, parliamentary institutions and legal protections.
The UK National Security Act does not criminalise criticism of China.
It does not criminalise academic cooperation with China generally.
It does not make every interaction with a Chinese citizen suspicious.
And today's CGTRI alert does not itself create a criminal offence for academics.
Those distinctions matter.
But there is another point worth making.
Democratic governments do not become immune from criticism merely because their national-security concerns are legitimate.
China can pose genuine intelligence risks to Britain.
That does not mean every legal response is automatically proportionate.
Likewise, Britain's intelligence assessment can be correct while individual academics remain entitled to due process.
Those propositions are perfectly compatible.
The timing makes the CGTRI case particularly significant
The government is simultaneously telling universities to become more careful about foreign research relationships while giving itself new legal tools for dealing with organisations it considers threats.
MI5 says more than 100 UK-linked academics have already been involved in CGTRI-funded projects.
And it acknowledges something important:
some academics may not have known that CGTRI was involved in the funding.
That matters enormously in a criminal-law context.
Universities operate through layers of grants, subcontractors, research partners and international institutions.
An academic may know the name of a university or laboratory without knowing the ultimate source of funding.
MI5 itself is telling universities to establish the ultimate funding source.
The security service is effectively saying:
You need to know who is ultimately paying for the research.
That is a reasonable national-security concern.
But it also places a substantial compliance burden on universities and researchers.
Britain is entering the same difficult territory it criticises elsewhere
This is ultimately the larger story.
Modern states increasingly face genuine threats from foreign intelligence services.
China conducts espionage.
Russia conducts espionage.
Western intelligence services conduct espionage too.
Universities are valuable targets because they contain advanced research, talented scientists and commercially important technology.
AI makes this even more significant.
A breakthrough in machine learning, cybersecurity, cryptography or autonomous systems can have both civilian and military applications.
MI5's warning therefore cannot simply be dismissed as paranoia.
But neither should national security become a magic phrase that ends the debate.
The important question is:
How much power should a democratic state acquire to protect itself from foreign intelligence — and what safeguards should accompany that power?
Britain's answer is becoming clearer.
It is moving toward a system in which foreign-intelligence offences cover not only traditional espionage but material benefits and information; and in which designated bodies can trigger additional offences covering assistance and, under defined circumstances, supportive expression.
That is a substantial expansion.
The uncomfortable question Britain should answer
There is a legitimate national-security argument for Britain to investigate foreign-funded academic research.
There is a legitimate argument for universities to know where their money comes from.
There is a legitimate argument for prosecuting people who knowingly assist foreign intelligence services.
The difficult question comes when the state moves from conduct to association, and from association to expression.
Once an organisation has been designated, Parliament has effectively created a legal environment in which supporting that organisation can potentially become criminal when the statutory purpose requirement is satisfied.
That deserves scrutiny regardless of whether the organisation is Chinese, Russian, Iranian, American or British.
Because national-security powers created for one adversary rarely remain conceptually limited to that adversary.
Today's justification may be Chinese espionage.
Tomorrow's case could involve another foreign state.
The precedent is the important part.
The lesson from Hong Kong should apply in both directions
Britain's criticism of Hong Kong should not be dismissed simply because Britain has passed tougher national-security laws of its own.
But neither should Britain's national-security legislation automatically be described as equivalent to Hong Kong's.
The proper test is more demanding.
Look at the actual statute.
Look at the powers it gives ministers.
Look at the role of Parliament.
Look at the standard applied by courts.
Look at the evidence required for conviction.
Look at the safeguards available to people affected by designation.
And then ask whether the balance between security and liberty remains proportionate.
That is the debate Britain should be having.
The MI5 warning about CGTRI makes that debate urgent because it demonstrates how quickly intelligence assessments can intersect with criminal law.
MI5 says more than 100 UK-linked academics have already been connected to research funded through the institute.
Some may have known exactly who was behind the money.
Others, MI5 says, may not have known at all.
And that is precisely why the new British national-security regime deserves close attention.
The question is no longer simply whether foreign intelligence services are trying to penetrate British institutions.
They almost certainly are.
The question is how far a democratic government should be allowed to go in stopping them — and what happens to individual liberty when national security becomes the justification for expanding that boundary.
Comments
Post a Comment