What’s New? September 2026

By Alex Heshmaty on September 17, 2026

AI bumper roundup

Lawyer struck off over AI misuse

The Solicitors Disciplinary Tribunal (SDT) decided to strike off registered foreign lawyer Abhishek Kumar for repeatedly submitting erroneous material generated by artificial intelligence (AI) in his defence pertaining to alleged misconduct. The SDT said that he had “embarked on the use of AI without giving any proper thought to the consequences” and failing “to take proper steps to ensure his work output was accurate or to satisfy himself that the electronic tools he was seeking to use were reliable.”

SRA warning notice

The Solicitors Regulation Authority (SRA) has issued guidance on the use and misuse of in the legal sector. The two key issues which it raises are: (i) the proliferation of “AI hallucinations” in court documents (a significant concern of lawyers) and (ii) the risks posed by uploading confidential client information to AI systems. The SRA says that solicitors “remain accountable for all work and advice delivered to clients, whether or not AI has been used in its preparation.”

AI firm wins court case

Garfield AI, the UK’s first regulated AI law firm, has achieved the first known instance of an AI law firm court victory in the world. Tamires Camal Taquidir, a freelancer who used Garfield to pursue £7,000 in unpaid HR service fees from a hospitality business, was represented by a human barrister, One Essex Court’s Dominic Li, who said the AI-drafted documents were “more than sufficient for the purposes of this trial”.

Dangers to justice

Dame Victoria Sharp, president of the King’s Bench Division, recently warned that increased reliance on AI tools by judges could lead to a “gradual drift” towards judicial standardisation based on potentially biased AI outputs, and – in the case of judges who treat AI as sources of judicial reasoning – effectively “outsourcing part of the judicial function” to AI companies. Concerns have also been raised about the dangers of AI in the executive branch of justice, after a police detective was reportedly under investigation for allegedly using an AI chatbot.

Negligence from using or not using AI

The UK Jurisdiction Taskforce (UKJT) released a Legal Statement on Liability for AI Harms, in which it suggests that lawyers could be sued for negligence if their avoidance of AI tools results in an unfavourable outcome for a client. Section 7 states that a legal professional “could be liable for failing to use AI in circumstances where a competent member of their profession would have done so.” However in the same section, it also states that a lawyer “may be found negligent for using AI inappropriately, for using an unsuitable model, for failing to conduct proper due diligence, or for failing to test AI or validate its outputs effectively.” Meanwhile, the Civil Justice Council (CJC), which is currently examining the use of AI by legal representatives for preparing court documents, has recently said that “no additional formal requirements relating specifically to AI are currently necessary.”

Potential legislation

The Artificial Superintelligence Bill was introduced by Alex Sobel MP on 8th September 2026, as a Private Members’ Bill, which aims to “prohibit the development, deployment and operation of artificial superintelligence systems”. Sobel has also backed calls from the Joint Committee on Human Rights (JCHR)- which he chairs – for a new AI Bill and AI-specific regulation. However, another legislative proposal by the MP and other colleagues, to introduce an AI “kill switch” has been rejected by the Cabinet Office, which oversees AI safety and research via its AI Security Institute (AISI).

Labelling AI

Article 50 of the EU AI Act, which came into force at the beginning of August 2026, requires the labelling of certain AI-generated content, so that anyone viewing the content is aware that it was created by AI. This follows a similar measure taken by New York State.

Legal AI

Legora recently announced a change to its fee model, moving from per seat to consumption based pricing, indicating a general shift which will likely put the kibosh on tokenmaxxing and reduce the volume of AI prompts in law firms. However, this is not putting off law firms keen to deploy their own AI products, including Ashurst and Shoosmiths.

AI Minister

The UK government has appointed its first ever Minister for Artificial Intelligence, Kanishka Narayan, under the new Burnham Labour government. One of his first duties was attending the launch of a new Frontier AI Research Lab by Imperial College and Thomson Reuters.

Legal miseducation

Universities are starting to grapple with the impact of AI on the way they teach and how to prepare students for a new world of work. The University of Chicago Law School has produced a policy statement entitled “Rethinking Legal Education in the AI Era” which outlines a range of measures including the decision to “prohibit the use of electronic devices such as laptops, tablets, and phones in the classroom.”

Political pressure

As we wrote about in this newsletter, the American government recently ordered Anthropic to revoke access to foreign nationals on the basis of national security, a move which provoked legal challenge from one US legal technology company. It’s also worth noting that the Trump administration is reportedly siding with OpenAI in its copyright battle with the New York Times. Meanwhile, President Trump’s Executive Order 14365 appears to be using the Federal Trade Commission (FTC) to challenge US States which are implementing legal restrictions on AI.

AI policy research

Various research projects are underway, looking at the impact of AI on legislation and regulation. This includes AI Scenarios 2030 (designed to help policymakers plan for the future of AI), an FCA review into the impact of AI on retail financial services, and ongoing work by the Digital Regulation Cooperation Forum (DRCF).

Online safety roundup

Meta law suit

Meta has agreed to pay $18 billion to settle a federal trial which included claims from 29 states that the social media giant violated the US Children’s Online Privacy Protection Act by knowingly collecting personal data from children without parental consent, and using the data to train generative AI. Although some commentators argue that this is social media’s “big tobacco” moment, the reality is that the cost of the settlement was dwarfed by a 1% rise in the value of shares as a result of stock market jitters being allayed in relation to the much larger penalties (some $1.4 trillion) which were sought in the trial. Aside from the money, which will be paid out over 10 years, the settlement also includes a commitment to certain measures designed to restrict the use of Meta services such as Instagram and Facebook for teenagers, including hiding likes and providing time limits, contingent on other social media apps deciding to implement similar changes. Although the settlement has been criticised as being rather paltry, it exceeds separate fines of $567 million and $375 million related to child safety failings. However, thousands of similar lawsuits are in the pipeline, which could gradually begin to eat into even the deep pockets of big tech.

Social media curfew

Shortly following its announcement of a social media ban for under-16s in the UK, the government raised the prospect of a “curfew” on use of social platforms for 16 and 17 year olds, alongside other plans to minimise addictive features of technology, including AI chatbots. However, it seems that the plans revolve around introducing certain restrictive default settings (eg which prevent the use of social media apps between midnight and 6am) which can simply be turned off by phone users, and have been described by some commentators as a “mildly annoying settings prompt with a government press release attached”. Culture secretary Lisa Nandy MP said the government will legislate to compel technology firms to prevent children from taking or sharing nude images on mobile devices, after their initial request to block explicit images at device level failed to garner results, and in light of growing online harms to children. Meanwhile, down under, the Australian government is doubling the fines for breaches of their social media ban, which around two thirds of children appear to be circumventing.

Regulatory setbacks

Despite continued efforts to regulate online platforms and recently publishing a Strategic Approach to AI in response to the government’s 2025 AI Opportunities Action Plan, Ofcom is facing challenges on multiple fronts, including legal action from Zuckerberg’s social media empire. Meta claims that the fees and penalties in relation to the Online Safety Act “should be based on the services being regulated in the countries they’re being regulated in” – which would in effect translate to a maximum fine of 10% of their UK revenue as opposed to worldwide revenue. Separately, Ofcom has been facing criticism recently over perceived failures to properly regulate the online space, and fellow UK technology regulator, the Information Commissioner’s Office (ICO), has been facing its own controversies of late.

Media regulation

The government has been consulting on the future of UK media, setting out its strategy in a green paper on a range of matters, including the consumption of broadcast news moving online. We delved into this topic in a recent piece: News broadcast regulation in an online age.https://www.bbc.co.uk/news/articles/c3j4jz8vpz1xo

Although some European countries are imposing their own social media restrictions for children, the EU has announced plans to implement harmonising legislation across member states.