iiinova — where your digital self keeps evolving

Zihao Li @zihaoli.bsky.social · 15/09/2026

We would love to see you there. Please help us sharing this invitation with colleagues and networks who may be interested. Suggestions for relevant empirical studies to add to the Wiki are also very welcome. Zoom Registration: lnkd.in/e3vG_TVe

lnkd.in

This link will take you to a page that’s not on LinkedIn

000

Zihao Li @zihaoli.bsky.social · 15/09/2026

000

Zihao Li @zihaoli.bsky.social · 15/09/2026

I’m delighted to share that we will be launching the 𝗗𝗮𝘁𝗮 𝗣𝗿𝗼𝘁𝗲𝗰𝘁𝗶𝗼𝗻 𝗘𝘃𝗶𝗱𝗲𝗻𝗰𝗲 𝗪𝗶𝗸𝗶 📅 Thursday, 𝟮𝟰 𝗦𝗲𝗽𝘁𝗲𝗺𝗯𝗲𝗿 🕐 1–2.30pm BST | Zoom Our team has reviewed more than 𝟭,𝟬𝟬𝟬 𝗽𝘂𝗯𝗹𝗶𝗰𝗮𝘁𝗶𝗼𝗻𝘀 covering the period from 1995 to January 2026, bringing together 𝟮𝟱𝟴 𝗲𝗻𝘁𝗿𝗶𝗲𝘀 of empirical evidence on the GDPR.

000

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 10/07/2026

🤖 AI-driven research and data protection: CREATe response to the ICO consultation Based on the UKRI Metascience AI project, led by Dr Zihao Li, the response examines how copyright and data protection law shape AI research — from data acquisition and model training to deployment and reuse.

111

Reposted by Zihao Li

Law and Humanities @law-and-humanities.bsky.social · 09/06/2026

New issue marks 100 years since first publication of Kafka's brilliant unfinished novel 'The Castle': guest editorial (Ekaterina Yahyaoui Krivenko) and all research articles, essays, arts and book reviews now available on the website: www.tandfonline.com/doi/full/10....

12217

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 08/05/2026

📣𝗡𝗲𝘄 𝗪𝗼𝗿𝗸𝗶𝗻𝗴 𝗣𝗮𝗽𝗲𝗿: #Accuracy #Paradox: Addressing Epistemic, Manipulative, and Societal Risks of #Hallucination in #AI #Governance by @zihaoli.bsky.social Li (CREATe), @weiweiyi.bsky.social Yi (CREATe), and 𝗝𝗶𝗮𝗵𝗼𝗻𝗴 𝗖𝗵𝗲𝗻 (University of Sheffield). Full blog 👉 www.create.ac.uk/working-pape... 1/2🧵

create.ac.uk

New working paper – Accuracy Paradox: Addressing Epistemic, Manipulative, and Societal Risks of Hallucination in AI Governance - CREATe

122

Zihao Li @zihaoli.bsky.social · 11/04/2026

👉 Hallucination should be understood not only as factual error, but also as misleading reasoning, false confidence, and biased framing 👉 Regulation (like EU AI Act) should go beyond the accuracy

000

Zihao Li @zihaoli.bsky.social · 11/04/2026

Key insights: 👉 Focusing on surface-level factual correctness hides deeper risks: oversimplification, loss of nuance, misrepresentation, and subtle distortions 👉 High-accuracy models can still produce outputs that sound right, feel authoritative, and yet influence decisions 8/n

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

We propose: ✔️ Moving from accuracy → epistemic reliability ✔️ Embedding uncertainty, verification, and contestability into AI systems ✔️ Preserving plurality and diversity of perspectives in AI outputs ✔️ Developing dynamic evaluation frameworks beyond static benchmarks (7/n)

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

We test current EU regulations, including the AI Act, GDPR and DSA, showing that existing frameworks overly focus on accuracy-related harms, while overlooking interactional, epistemic, and behavioural risks that emerge in GenAI. (6/n)

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

We develop a taxonomy of hallucinations, distinguishing between different types risks: - Factual hallucinations - Consistency hallucinations - Sycophantic hallucinations - Contextual / interactional distortions - Epistemic, oversimplifiedand normative hallucinations -etc. (5/n)

110

Zihao Li @zihaoli.bsky.social · 11/04/2026

We conceptualise the “accuracy paradox”: systems overly optimised for accuracy may simultaneously - reinforce epistemic convergence - reduce plurality of perspectives - increase susceptibility to persuasive or manipulative outputs (4/n)

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

We argue that overreliance on accuracy as the primary benchmark for AI governance is not only incomplete, but may also be counterproductive. Focusing narrowly on accuracy obscuring deeper epistemic, behavioural, and societal harms that emerge in dynamic human–AI interaction (3/n)

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

GenAI has intensified concerns around AI hallucination, regulatory and technical debates have largely converged on one central objective: 𝑖𝑚𝑝𝑟𝑜𝑣𝑖𝑛𝑔 𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦. Our paper challenges this assumption. (2/n)

100

Zihao Li @zihaoli.bsky.social · 11/04/2026

🚨𝗡𝗲𝘄 𝗣𝗮𝗽𝗲𝗿 𝐀𝐥𝐭𝐞𝐫🚨 With @weiweiyi.bsky.social and Jiahong Chen (@sheffielduni.bsky.social), our paper Accuracy paradox: Addressing epistemic, manipulative, and societal risks of hallucination in AI governance, pblished in 𝐂𝐨𝐦𝐩𝐮𝐭𝐞𝐫 𝐋𝐚𝐰 & 𝐒𝐞𝐜𝐮𝐫𝐢𝐭𝐲 𝐑𝐞𝐯𝐢𝐞𝐰 doi.org/10.1016/j.cl...

doi.org

Redirecting

101

Reposted by Zihao Li

Lilian Edwards @lilianedwards.bsky.social · 06/03/2026

Wooo wooo ! you can now pre order and request inspection copies of Law Policy and the Internet, 2nd edn! w me, @mooseabyte.bsky.social and @cgoanta.bsky.social eds and an all star line up! Lots of AI but guaranteed no AI slop! Out August! www.bloomsbury.com/uk/law-polic...

bloomsbury.com

Law, Policy, and the Internet

This innovative textbook edited by international experts in technology law and policy provides a comprehensive and accessible guide into the fast-evolving lands…

32510

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 19/02/2026

Workshop Recap - Revisiting #AI #Justice in #China⚖️🤖 On 28 January 2026, as part of CREATe's 𝘈𝘶𝘵𝘰𝘮𝘢𝘵𝘪𝘰𝘯, 𝘋𝘦𝘤𝘦𝘯𝘵𝘳𝘢𝘭𝘪𝘴𝘢𝘵𝘪𝘰𝘯 𝘢𝘯𝘥 𝘗𝘭𝘢𝘵𝘧𝘰𝘳𝘮 theme, @zihaoli.bsky.social hosted a workshop with Weiwei Shen (China University of Political Science and Law) on the rapid rise of AI justice in China. 1/4🧵

create.ac.uk

Revisiting AI Justice in China: Rationale and Risks behind the Progressive Movement - CREATe

111

Zihao Li @zihaoli.bsky.social · 20/01/2026

🗓 𝑺𝒖𝒃𝒎𝒊𝒔𝒔𝒊𝒐𝒏 𝒅𝒆𝒂𝒅𝒍𝒊𝒏𝒆: 15 𝑺𝒆𝒑𝒕𝒆𝒎𝒃𝒆𝒓 2026 CfP: www.sciencedirect.com/special-issu... If you know someone who may be interested, please feel free to share. If you have any questions about the Special Issue, you are very welcome to contact Wenlong and me directly.

000

Zihao Li @zihaoli.bsky.social · 20/01/2026

The scope of this Special Issue should be broadly defined. We welcome comparative, empirical, doctrinal, and interdisciplinary contributions, covering one or more Asian jurisdictions, as well as global comparative perspectives. 3/n

100

Zihao Li @zihaoli.bsky.social · 20/01/2026

A lot of things are happening across Asia in the field of AI regulation. From emerging GenAI laws, to AI watermarking, transparency, and content governance, Asian jurisdictions are developing regulatory approaches that go well beyond a simple extension of the Brussels Effect. 2/n

100

Zihao Li @zihaoli.bsky.social · 20/01/2026

📣 𝐂𝐚𝐥𝐥 𝐟𝐨𝐫 𝐩𝐚𝐩𝐞𝐫𝐬: 𝐀𝐈 𝐑𝐞𝐠𝐮𝐥𝐚𝐭𝐨𝐨𝐧 𝐢𝐧 𝐀𝐬𝐢𝐚 Together with @wenlongli.bsky.social, we are guest editing a Special Issue in 𝑪𝒐𝒎𝒑𝒖𝒕𝒆𝒓 𝑳𝒂𝒘 & 𝑺𝒆𝒄𝒖𝒓𝒊𝒕𝒚 𝑹𝒆𝒗𝒊𝒆𝒘 on AI Regulation in Asia: Emerging Pathways, Divergent Models, and Global Implications. CfP: lnkd.in/gRgGXrv3 1/n

lnkd.in

This link will take you to a page that’s not on LinkedIn

100

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 24/11/2025

🚨 𝗡𝗲𝘄 𝗪𝗼𝗿𝗸𝗶𝗻𝗴 𝗣𝗮𝗽𝗲𝗿 🚨 How should we regulate #DarkPatterns effectively? ⚖️🖥️ In our latest entry in the 2025 Working Paper Series, @weiweiyi.bsky.social and @zihaoli.bsky.social (CREATe, University of Glasgow) offer a systematic review on dark patterns regulation across Law and HCI

122

Zihao Li @zihaoli.bsky.social · 11/11/2025

We propose a paradigm shift toward “Diligent Design”, a proactive, design-embedded regulatory model that integrates legal and technical accountability. A call for collaboration between lawyers, policymakers, and designers. (6/6)

010

Zihao Li @zihaoli.bsky.social · 11/11/2025

We also explain why existing frameworks, in data protection, consumer, and competition law, fail to address the problem effectively. Regulators are trapped in a reactive loop of “proving harm.” (5/n)

110

Zihao Li @zihaoli.bsky.social · 11/11/2025

Our review bridges that gap. We identify 5 Root Problems driving dark patterns from power asymmetry to attention-seeking business models. And 3 Layers of Harm they cause: Individuals, Markets, and Society itself. (4/n)

120

Zihao Li @zihaoli.bsky.social · 11/11/2025

We reviewed 65 interdisciplinary studies across Law and HCI. The finding? Scholarship is dangerously fragmented. HCI builds taxonomies. Law assesses new/failed regulations. The two disciplines don’t speak the same language — and rarely speak to each other. (3/n)

110

Zihao Li @zihaoli.bsky.social · 11/11/2025

Why is the regulation of dark patterns, manipulative interface designs that steer users into unwanted actions, still so ineffective? 🤔 Our paper systematically reviews how Law & HCI have studied dark patterns and uncovers the root causes of regulatory failure. (2/n)

110

Zihao Li @zihaoli.bsky.social · 11/11/2025

🚨 NEW PUBLICATION ALERT 🚨 Thrilled to share our latest paper (with @weiweiyi.bsky.social): “Mapping the Scholarship of the Regulation of Dark Patterns” now published in Computer Law & Security Review. Open Access link: doi.org/10.1016/j.cl...

doi.org

Redirecting

110

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 05/09/2025

CREATe's brilliant minds are on the move! This week, our researchers, @ulafurgal.bsky.social and @zihaoli.bsky.social, took the stage at the #SLS annual conference 2026 by The Society of Legal Scholars in Leeds, showcasing their cutting-edge work.👏 We hope the audience enjoyed the show by CREATe!

011

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 29/07/2025

3/3 Transparency requirements, opt-outs, the exception for text and data mining, and equitable remuneration for creatives are among the key issues we analysed in our response to the UK copyright & AI consultation, which has been published in GRUR International (academic.oup.com/grurint/adva...)

academic.oup.com

Abstract. The interface between copyright law and artificial intelligence (AI) is currently the object of global attention. The UK government’s recent publ

022

Zihao Li @zihaoli.bsky.social · 23/04/2025

We hope this paper serves as a basis for a meaningful evaluation of the GDPR, and provides empirical evidence that may assist other jurisdictions in assessing the effectiveness of transplanting GDPR-inspired rules into their own data protection frameworks.

000

Zihao Li @zihaoli.bsky.social · 23/04/2025

Our paper makes a three-fold contribution (1) a systematic mapping & critical analysis of the empirical evidence (2) a taxonomy of evidence for GDPR evaluation & review (3) reflections on the methodological deficits for more coordinated future research agenda. (5/n)

100

Zihao Li @zihaoli.bsky.social · 23/04/2025

Particularly in recent evaluations, at least three fallacies are present: 1 binary success/failure frame, 2 disproportional focus on GAFAM & 3 overgeneralisation. (4/n)

100

Zihao Li @zihaoli.bsky.social · 23/04/2025

By categorising empirical evidence into four thematic clusters—Awareness and Trust, Operational Performance, Ripple Effect, and Normative Clarity—we provides a structured, multidimensional understanding of GDPR implementation and implications. (3/n)

100

Zihao Li @zihaoli.bsky.social · 23/04/2025

We have conducted a systematic review and reflection of all empirical literature on EU data protection law, including both the Directive and the GDPR, over the past 25 years (1995–2022) (2/n)

100

Zihao Li @zihaoli.bsky.social · 23/04/2025

As the GDPR enters its seventh year of implementation, and with recent calls to reduce GDPR-related bureaucratic burdens, it may be the right time to re-examine the real-world impact and future direction of this landmark legislation. (1/n)

100

Zihao Li @zihaoli.bsky.social · 23/04/2025

🚨NEW ARTICLE ALERT🚨 I am thrilled that our article (@wenlongli.bsky.social @wenkaili.bsky.social Yuemming Zhang @aolanli.bsky.social ) ‘Mapping the empirical literature of the GDPR's (In-)effectiveness: A systematic review’ published by Computer Law & Security Review. OA: doi.org/10.1016/j.cl...

243

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 16/04/2025

🚨 New CREATe Working Paper on the GDPR 🚨 ‘Mapping the Empirical Literature of the GDPR’s (In-)Effectiveness’ By WenlongLi, @zihaoli.bsky.social (Zihao Li), @Wenkai Li, Yueming Zhang & Aolan Li Read here 👉 www.create.ac.uk/blog/2025/04...

111

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 04/04/2025

Some of our colleagues are currently presenting their latest work at the BILETA Conference in @goldsmithsuol.bsky.social 🧵👇

121

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 04/04/2025

Weiwei Yi on dark monetisation design patterns in video games 🎮 Visit our website to discover more about our research: www.create.ac.uk

create.ac.uk

CREATe – Centre for Regulation of the Creative Economy

011

Zihao Li @zihaoli.bsky.social · 26/03/2025

New Working Paper - The Decentralized and Automated Economy: A research roadmap for the medium-term future - www.create.ac.uk/blog/2025/03...

create.ac.uk

New Working Paper – The Decentralized and Automated Economy: A research roadmap for the medium-term future – CREATe

011

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 28/02/2025

🚨 Our response to the UK Government’s consultation on AI and copyright 🚨 🎨 🤖 Is it possible to adjust the legal framework to support both the creative industries and AI development? Response here 👉 www.create.ac.uk/blog/2025/02...

011

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 27/01/2025

📢 Job Alert! The University of Glasgow is hiring a Lecturer in Competition Law (Research & Teaching). Join a vibrant team, teach innovative courses, and contribute to world-class research with CREATe. Closing: 24 Feb 2025. Check our blog for more info: www.create.ac.uk/blog/2025/01...

create.ac.uk

Vacancy: Lecturer in Competition Law (Research & Teaching Track) – CREATe

054

Reposted by Zihao Li

CREATe @create-glasgow.bsky.social · 16/02/2024

We are thrilled to announce the @create-glasgow.bsky.social relaunch event on 11 March at the ARC. Come along to learn more about our work on creativity, markets & technology, new research themes & join us for a networking reception with a Scottish twist. Register: www.create.ac.uk/blog/2024/02...

1102