MLex: The Chinese operator of the DeepSeek generative AI model has been questioned by the Irish data regulator about how it processes the data of people in Ireland, the watchdog has confirmed to MLex. The move follows a similar request from Italy’s data protection authority as pressure grows from consumer groups concerned about what data is collected, how it is processed and where it is stored.

MLex: DeepSeek is being investigated by the Information Commissioner’s Office (ICO) over data protection concerns. The ICO joins regulators in a growing list of countries seeking answers from the Chinese AI developer, including Germany, South Korea, Italy, Australia and India.

This Practice Note provides a summary of UK law as it applies to the use of deepfakes. A deepfake is a form of audiovisual content that has been generated or manipulated using artificial intelligence (AI) that misrepresents someone or something. The Practice Note covers: what a deepfake is; how deepfakes work; and uses and applications of deepfakes—such as for entertainment, parody, political satire and healthcare. It also looks at the application of UK law to deepfakes, what social media platforms and search engines are doing to counter the issues created by deepfakes, and future technological controls including problem areas concerning technological and legislative or common law controls. What is a deepfake? The term deepfake is a portmanteau of ‘deep learning’ and ‘fake’. In its Deepfake Defences Discussion Paper, Ofcom defines a deepfake as ‘a form of audio-visual content that has been generated or manipulated using AI that misrepresents someone or something’. Deepfakes often feature individuals in the entertainment, fashion, or sports sectors. Increasingly, major industry players such as Disney are actively developing their own variants. Deepfakes

Law360: In a high-stakes arbitration, lawyers for one of the companies present what they say is surveillance video of a bribe being accepted by its opponent's president. They argue the video presents incontrovertible evidence that the case should be decided in their client's favor—and a tribunal might be inclined to agree. But what if it turns out that the video is a fake, generated by artificial intelligence?

A security where the amount payable on maturity, or any other occasion when the security can be redeemed will or may exceed the issue price by more than 0.5% for each year in the redemption period, up to a maximum of 30 years. See: SAIM3020.

TMT analysis: DeepSeek AI is an advanced artificial intelligence model developed by DeepSeek, a Chinese company. Their mission is ‘dedicated to pushing the boundaries of artificial intelligence to create solutions that empower businesses and enhance human capabilities, (…) [and] to make advanced AI technology accessible, ethical, and impactful’. Founded in 2023, DeepSeek quickly emerged as a major player in AI with its free open-source language models. It recently launched two advanced models, V3 for general applications like conversational AI, and R1 for tasks involving reasoning, such as programming and math problems. R1 has brought significant media attention to DeepSeek for its cost-effective AI solutions compared to leading US models. Written by Thomas Dubuisson, senior associate at CMS Belgium.

Jane Anderson, senior associate, Amy Au, associate, and Jason Watkins, trainee, at Irwin Mitchell, review the Fiona Harvey versus Netflix defamation claim from an English law standpoint.

Law360, London: Deezer has hit a marketing software business with a claim of almost €12m at a London court, alleging that the technology company negligently handled users' data that was eventually hacked and sold on the dark web.

A portmanteau term covering both libel and slander.

The tort of defamation is governed by a mixture of statute and common law. The relevant statutory law is contained in: • the Defamation Act 1952 (DA 1952) • the Defamation Act 1996 (DeA 1996) • the Defamation Act 2013 (DA 2013) There is no statutory definition of what is defamatory. The conventional common law test is that the imputation must tend to lower a claimant in the estimation of right-thinking members of society generally. However, DA 2013, s 1 introduced a threshold requirement that a statement is not defamatory unless its publication has caused or is likely to cause serious harm to the reputation of the claimant, giving statutory effect to the common law threshold tests in Thornton v Telegraph Media Group and Jameel (Yousef) v Dow Jones & Co but raising the threshold bar from the common law requirement of 'substantial' harm to 'serious' harm. In Lachaux, the Supreme Court held that DA 2013, s 1 also introduced a requirement that a claimant must demonstrate as a fact that they have suffered, or are likely to suffer

This Practice Note examines the role of publication in an action for defamation or malicious falsehood. It reviews key case law and legislation, and considers who may be primarily and secondarily responsible for publication, what constitutes publication, the nature of publication on the internet and jurisdictional issues. Publication as an ingredient of the tort of defamation It is an essential ingredient of an action for libel or slander that there has been communication of a defamatory statement to a person other than the claimant. For a written publication to occur, the words in question must be read and understood by a third party. For an oral publication to occur, the words must be apprehended and understood by a third party. In practice a claimant brings one action in relation to multiple publications. Each communication is, in principle, a separate and distinct publication giving rise to a separate cause of action, although section 8 of the Defamation Act 2013 (DA 2013) moderated this principle for the purposes of the limitation of the action (see below). Further reading You

This Practice Note covers defamation in the context of social media. It examines how liability for defamation may arise for individuals, internet service providers (ISPs), website operators and employers and explores defences under the Defamation Act 1996 (DeA 1996), the Electronic Commerce (EC Directive) Regulations 2002 (E-Commerce Regulations 2002), SI 2002/2013, the Defamation Act 2013 (DA 2013) and the Defamation (Operators of Websites) Regulations 2013 (DOW Regulations 2013), SI 2013/3028. The internet and, in particular, social media is a high risk area for defamation litigation. Content posted on blogs, internet forums, online newspaper comment sections and well known social networking sites such as Facebook and X (formerly Twitter) can be inaccurate and vicious in character. Anonymity encourages some individuals to dispense with the usual restraints that they might apply to other forms of publication. For a description of the popular social media websites, how they are used by businesses and associated legal issues, see Practice Note: Introduction to social media. Liability for defamatory content posted on social media is not, in principle, restricted to the author (who in