Introduction
Imagine your product team ran a UK FTO analysis 18 months ago for an AI-implemented feature. The opinion came back clean. Under the then-current Aerotel test, the relevant class of software-implemented inventions was excluded from patentability at the threshold stage — they wouldn’t grant, so they couldn’t block you. That analysis was accurate for its time.
On 11 February 2026, the UK Supreme Court abandoned the Aerotel test in Emotional Perception AI v UKIPO [2026] UKSC 3. The UK now assesses computer-implemented inventions under the EPO’s ‘any hardware’ approach. A broader class of AI and software-related inventions can now be patented in the UK. And the FTO analysis your team relied on was built on assumptions that no longer apply.
This isn’t an abstract legal development. It has direct, practical consequences for any product team or IP function relying on UK FTO clearance for AI-implemented technology. This article explains what changed, why it matters for FTO, and what to do now.
What the UK Supreme Court Actually Decided
The Aerotel test and its FTO significance: Since 2006, the UK courts applied a four-step test from Aerotel v Telkom to assess whether a computer-implemented invention was excluded from patentability as a ‘program for a computer as such’ or a ‘method for performing a mental act as such’. The test required identification of the actual contribution the invention made, then assessment of whether that contribution fell solely within excluded subject matter. In practice, Aerotel excluded a wide range of software-implemented inventions from UK patentability — including many AI-related inventions — at the threshold stage, before novelty or inventive step were even assessed. For FTO purposes, this was significant: patents that couldn’t grant couldn’t block.
What Emotional Perception AI v UKIPO decided: The Supreme Court held that the Aerotel test was wrong and should no longer be applied. Instead, the UKIPO must assess computer-implemented inventions using the EPO’s approach — specifically the ‘any hardware’ or ‘technical character’ test, sometimes called the intermediate step approach. An invention that involves any hardware — even a general-purpose computer — is no longer automatically excluded at the threshold stage as a ‘computer program as such’. The exclusion applies only to the program as such — an abstract program divorced from any technical implementation. In practice, very few genuine AI-implemented inventions are abstract in that sense.
The new standard in practice: Under the EPO approach, a computer-implemented invention passes the patentability threshold if it involves a ‘technical means’ — which can be satisfied by running on standard hardware. The substantive question then shifts to novelty and inventive step, assessed in the normal way. For AI inventions, this is a materially lower bar to patentability than Aerotel created. More AI and software patents will reach grant stage in the UK. The FTO landscape for AI-implemented features has widened.
DATA NOTE: The UKIPO immediately withdrew its AI examination guidelines following the February 2026 ruling, signalling that prior guidance on AI patent prosecution no longer reflects the applicable legal standard. New UKIPO guidance on AI patentability is expected in late 2026 — but the Supreme Court ruling itself is effective immediately.
Why Your Existing UK FTO May No Longer Hold
If your team has UK FTO analyses covering AI or software-implemented features, the key question is whether those analyses relied — explicitly or implicitly — on the Aerotel threshold exclusion as part of the clearance reasoning. Our analysis of the 5 most common pitfalls in FTO analyses highlights that one of the most frequent and consequential errors is relying on assumed patent invalidity or non-grantability without directly verifying current legal status. Post-February 2026, any FTO clearance that was grounded in ‘these patents wouldn’t grant under Aerotel’ is exactly that kind of assumption — and it no longer holds.
The pending application risk: The most significant near-term FTO risk is competitor patent applications that were filed before February 2026 but not yet granted. Under Aerotel, many of these applications would have been refused at the threshold stage. Under the new standard, the UKIPO must reassess them under the EPO approach. Applications covering AI-implemented features in your technology space that were borderline under Aerotel — and may have been excluded from your FTO scope on the basis that they wouldn’t grant — may now grant. The FTO scope that excluded them was correct for the old legal framework. It is not correct for the new one.
How the landscape widens: The Aerotel threshold excluded not just obviously abstract programs, but a wide range of genuinely technical AI and software implementations that the EPO would have allowed through. Methods of training neural networks, AI-implemented signal processing, computer vision algorithms running on standard hardware, natural language processing implemented in specific technical contexts — all of these categories faced a higher Aerotel barrier than they face under the new EPO-comparable standard. The UK AI and software patent landscape going forward will progressively include patents that would not have been granted before February 2026.
What the New Standard Means in Practice
More AI inventions can reach grant: The immediate practical consequence is that UKIPO examination of AI and software patent applications is now conducted under a more permissive threshold standard. Applications already in prosecution that were at risk of Aerotel refusal are being reassessed. New applications covering AI-implemented technology will be examined from the outset under the EPO-comparable approach. Over the next 12–24 months, the volume of granted UK patents covering AI-implemented technology will increase materially.
Transition uncertainty — the intermediate step: The EPO approach requires what is sometimes called an ‘intermediate step’ or ‘any hardware’ analysis: the examiner identifies which features of the claim contribute to technical character, then assesses novelty and inventive step only for those features. The UKIPO has no history of applying this methodology — Aerotel replaced it in 2006. There is currently no UK binding case law on how the intermediate step works in practice. Until a body of UK case law develops, there is genuine uncertainty about how the UKIPO will apply the test. For FTO purposes, the conservative strategy is to assume that borderline pending applications in your technology space may now grant under the new standard, and to scope FTO searches accordingly.
“The FTO analyses done before February 2026 were not wrong. They were accurate for the legal framework that existed when they were written. The framework has changed. The question now is whether the product you are launching today is still cleared under the framework that applies to it.”
Five Actions to Take Now
- Audit UK FTO analyses from before February 2026 covering AI or software features. Identify which analyses relied on Aerotel threshold exclusion as part of the clearance reasoning — explicitly or by omitting a class of patents on the basis that they wouldn’t grant. Those are the analyses that need reassessment first.
- Review competitor pending applications filed before the ruling. Patent applications filed by competitors before February 2026 covering AI-implemented technology in your space may now be proceeding through UKIPO examination under the new standard. These applications were potentially excluded from FTO scope on Aerotel grounds. Check whether any are now progressing toward grant.
- Update FTO scope for any UK market searches commissioned from now. Any FTO search for UK market clearance on AI or software-implemented technology must be scoped to the EPO-comparable standard — treating any hardware implementation as a potential basis for patentability. The old Aerotel scope is no longer the right framework.
- Monitor UKIPO examination outcomes as the new standard is applied. The first wave of UKIPO examination decisions under the new standard will clarify how the intermediate step is being applied in practice. This case law will materially affect how FTO scope should be calibrated going forward. Track UKIPO decisions in your technology space over the next 12 months.
- Consider refiling previously refused UKIPO applications in your own portfolio. If your organisation had AI or software patent applications refused under Aerotel that would likely pass the EPO intermediate step test, the ruling has effectively reopened those applications for reconsideration. Seek advice on whether refiling — with updated claim language where necessary — is appropriate. Our guide on FTO search best practices covers why FTO strategy and your own patent filing position need to be assessed together — because the gaps in your own portfolio are just as important as the risks in your competitors’.
How Our FTO Service Covers the Post-Aerotel UK Landscape
Our freedom to operate service covers UKIPO-granted and pending patents under the post-Aerotel EPO-comparable standard — including review of competitor pending applications that were filed before the ruling and may now be progressing toward grant under the new framework. For AI product teams with existing UK FTO analyses, we provide targeted reassessment reports identifying which conclusions remain valid under the new standard and which need updating. For new FTO commissions covering AI or software-implemented technology in the UK, we scope searches from the outset under the EPO-comparable threshold standard, not the Aerotel framework that no longer applies.
Your UK FTO for AI features may have been done under the Aerotel framework that was abandoned in February 2026. Our service covers the post-ruling UK landscape — including pending applications that may now be progressing toward grant under the new standard. → Contact Us
Conclusion: The Takeaway
The Aerotel test governed UK AI and software patent practice for nearly 20 years. Its abandonment in February 2026 is the most significant change to UK patent practice in a generation — not because it immediately invalidates existing patents, but because it widens the class of AI and software inventions that can now be granted in the UK going forward.
FTO analyses done before that date were accurate for the legal framework that existed when they were written. The question every product and IP team should be asking now is not whether those analyses were good work. They were. The question is whether the product launching today — in a UK market where the patentability threshold for AI-implemented technology has materially shifted — is still covered by those analyses. In most cases, the honest answer is: not without a reassessment.
