Is your bot talking $£!? about me again?

Defamation by AI: who is liable and what defences are available?

This podcast is intended as an introduction to issues that arise when an AI bot creates defamatory content. For detailed commentary on this specialist area of law, see: Gatley on Libel and Slander (12th Ed, 2017) and Duncan and Neill on Defamation (4th Ed, 2015 – with new addition forthcoming). For an overview, see our chapter on ‘Liability for Economic Harm’ in The Law of Artificial Intelligence (2020, forthcoming).

Cases relevant to auto-generated content include: 
  • Bunt v Tilly [2006] EWHC 407 (QB)
  • Metropolitan International Schools Ltd (trading as Skillstrain and/or Train2Game) v Designtechnica Corpn (trading as Digital Trends) and others [2009] EWHC 1765 (QB)
  • Tamiz v Google Inc. [2013] EWCA Civ 68 CA
For other jurisdictions, see e.g. Defteros v Google LLC [2020] VSC 219 at [40], in which Richards J summarised the Australian position as follows: “The Google search engine … is not a passive tool. It is designed by humans who work for Google to operate in the way that it does, and in such a way that identified objectionable content can be removed, by human intervention, from the search results that Google displays to a user.”  For Hong Kong, see e.g. Yeung v Google Inc. [2014] HKCFI 1404; Oriental Press Group Ltd v Fevaworks Solutions Ltd [2013] HKFCA 47 (especially [76] for a test endorsed by the authors of Gatley).

On the contradictory positions taken by search engines worldwide, see, e.g., Sookman, “Is Google a publisher according to Google? The Google v Equustek and Duffy cases”, C.T.L.R. 2018, 24(1).
(c) 2020, Matthew Lavy & Iain Munro