This monthly publication is designed for the wider tech‑law community, offering concise, accessible insights into the issues shaping our sector.
Each edition will include short pieces on topics such as:
- Recent deals and market developments
- Key legal updates in technology and AI
- Emerging trends across the tech industry
We welcome contributions from trainees, junior lawyers, and anyone across the tech‑law community who is keen to get involved.
For more information or to express interest in contributing, please contact hello@scl.org.
Ask Jeeves is dead – long live… TBC?
On May the 1st the website Ask.com closed down. Originally launched in 1997[1] as AskJeeves, its point of difference was a web search platform with results derived from actual questions – “is there life on Mars?” – rather than trying to obtain results using the dark arts of Boolean terms other search engines preferred. The website’s logo of the eponymous Jeeves, a cartoon butler in the fine PG Wodehouse and Agatha Christie tradition, affirmed this was a more welcoming and friendly website than might be expected by the online novice.
AskJeeves was an example of a more human approach to the internet, demonstrating the transition from the niche pursuit of Web 1.0 to the exponential growth in users and wider societal embedding of Web 2.0. This is the era where household-name brands and organisations were tentatively setting up digital shop (see Welcome.Jpeg’s recent survey of early high-fashion brand websites for some intriguing early steps[2]).
The growth of Web 2.0 created significant legal and societal challenges: copyright infringement due to filesharing sites, easy access to age-restricted content and social media-enabled hate crime, to name but three. Legislators were generally slow to catch up with these challenges, perhaps engaging only when companies and campaign groups outlined the impact of the harm from such activities. AskJeeves started in an era when it might happily share illegal music download websites, before ISP blocks and the likes of DMCA warnings.
As leading search engines developed their search algorithms and expanded into other product areas others like AskJeeves – simplified to Ask.com in 2006 – were left behind. The irony with AskJeeves finally closing is that, if it had launched today, it would be joining a market where search using natural language is gaining ground. Large Language Model AI tools are starting to look like potential competitors to established search platforms. Ofcom states one of the most prominent generative AI sites received 1.8 billion UK visits in the first 8 months of 2025 alone[3]. The trend for AI-powered search isn’t restricted to the “pure” AI platforms: numerous search engines are using this to augment traditional search, with an estimated 30% of search results now showing AI summaries[4]. However, the ascendancy of AI search isn’t inevitable; one well-established search engine has reported a three-fold increase in traffic with its ‘No AI’ search policy[5].
This shift in search, and the integration of AI into the internet, means the legal landscape will need to respond to these developments. With the ability to create increasingly convincing content, such as deepfakes and AI-generated text indistinguishable from human output, it could be argued some parts of the Online Safety Act 2023 are already outdated. Getty Images has been granted appeal to the Court of Appeal in its claim against Stability AI[6]. It is not impossible the long-standing section 9 provision in the Copyright, Designs and Patents Act 1988 on authorship could be scrutinised, given the debate on what protections – if any – should reside in AI-generated works.
Whilst AskJeeves may be out of contention, its spirit is alive and well in AI-powered search, and poses questions on the future of the internet – and the laws and bodies that aim to regulate it.
Dale Cornish is a Trainee Solicitor at Mayer Brown.
[1] https://searchenginewatch.com/1997/08/04/the-search-engine-report-august-5-1997-number-9/
[2] https://www.instagram.com/p/DYA58jNFACK/
[3] https://www.ofcom.org.uk/media-use-and-attitudes/online-habits/from-apps-to-ai-search-how-the-uk-goes-online-in-2025
[4] https://www.ofcom.org.uk/media-use-and-attitudes/online-habits/from-apps-to-ai-search-how-the-uk-goes-online-in-2025
[5] https://www.the-independent.com/tech/duckduckgo-ai-google-search-b2986989.html
[6] Getty Images (US), Inc & Ors v Stability AI Ltd
Disclosure or Understanding? AI Explainability and Trade Secrets
The concept of explainability has become a central feature of contemporary AI regulation. At its broadest, explainability refers to the ability to explain how an AI system reaches a particular outcome. Yet the growing emphasis on explainability raises an interesting question: how much information must be provided?
This question matters for several reasons. Individuals may wish to understand why a decision affecting them was made, regulators require explanations to assess compliance with legal obligations, and organisations may wish to demonstrate that their systems operate fairly, reliably and as intended. Explainability is therefore not simply a question of technical design. It is ultimately concerned with whether AI-driven decisions can, or should, be made intelligible to those affected by them.
Recent European developments suggest that questions of explainability cannot be considered separately from questions of confidentiality. Underlying these developments is a fundamental question: what kind of information genuinely functions as an explanation?
Meaningful Information
One example is Dun & Bradstreet Austria GmbH v D, in which the Court of Justice of the European Union (CJEU) considered an individual’s right to information about an automated decision-making process under the GDPR. The case concerned Article 15(1)(h) GDPR, which gives data subjects a right to obtain “meaningful information about the logic involved” in certain forms of automated decision-making. The Court held that individuals must receive meaningful information about the procedures and principles used in reaching a decision. The explanation provided must enable the person concerned to understand how their personal data were used and how the outcome was reached. Significantly, the Court rejected the idea that an organisation can avoid explanation simply by pointing to the complexity of the system involved.
At the same time, the Court recognised that requests for information may engage competing interests, including trade secret protection. If meaningful information must be provided, the practical question is how much information is enough.
The difficulty is that not every explanation serves the same purpose. A highly technical description of a model’s operation may reveal a great deal about the system while providing very little practical assistance to the individual receiving it. Similarly, a simplified explanation may be accessible and informative without revealing much about the underlying technology. The judgment offers limited guidance on where that line should be drawn. Instead, it leaves room for competing interests to be balanced. In doing so, it acknowledges that explainability and confidentiality are not necessarily mutually exclusive concepts.
Confidential Business Information
The same issue can be seen in the EU AI Act. The Act introduces a range of transparency and documentation obligations. Interestingly, the Regulation contains repeated references to the protection of trade secrets and “confidential business information”. This is significant because it suggests that the legislation does not proceed on the assumption that transparency requires unrestricted access to AI systems. Instead, the AI Act appears to distinguish between explaining a system and disclosing every aspect of it.
That distinction matters because the information capable of explaining a system may overlap with information regarded as commercially sensitive. In some circumstances, aspects of model design, training processes and evaluation methods may attract trade secret protection, particularly where they form part of a business’s competitive advantage. Dun & Bradstreet Austria makes clear that businesses cannot avoid transparency obligations simply by describing information as confidential. The existence of a trade secret does not, however, end the analysis.
Trade secret protection is significant not because it defeats explainability, but because it exposes a tension at the heart of it. Information about how a system has been designed, trained or evaluated may assist understanding, while also forming part of a business’s competitive advantage. Trade secret claims therefore force a closer examination of what explainability requires. This tension highlights a more fundamental question: is explainability concerned with access to information, or with enabling genuine understanding?
Discussions about AI transparency often focus on access to information. Increasingly, however, the more difficult issue is whether the information provided genuinely functions as an explanation. There is a distinction to be drawn between disclosing information and making a decision intelligible. The former is relatively easy to identify, and the latter is much harder to assess. The difficulty is not simply that businesses wish to keep certain information confidential. It is that the information most capable of explaining how a system operates may, in many cases, be the very information that businesses are most reluctant to disclose.
Looking Ahead
Both Dun & Bradstreet Austria and the AI Act suggest that European law is beginning to contemplate this reality. The present approach acknowledges the need for a better collective understanding of AI systems without necessarily requiring complete visibility of the underlying technology.
The requirement is clear: “meaningful information about the logic involved” or a “clear and meaningful explanation” of the role played by an AI system in coming to a particular decision. What remains much less clear is when those standards have been satisfied in practice. Neither Dun & Bradstreet Austria nor the AI Act provide a definitive answer. Both, however, proceed on the basis that explaining a system is not the same thing as opening it up for inspection.
Rebekah Moore is a Trainee Solicitor at Dentons.
Cartels by Code: The Antitrust Risks of Algorithmic Pricing
Since as early as 2018, the UK Competition and Markets Authority (the “CMA”) has probed the potential adverse effects of algorithmic pricing, warning that the use of such systems in many businesses’ operations may hinder fair competition to the detriment of consumers. With recent investigatory action in the UK echoing an increased interest in such tools from competition authorities across the globe, it is clear that algorithmic pricing presents a palpable risk to competitive markets.
What is Algorithmic Pricing?
Algorithmic pricing refers to software that continuously monitors data on market conditions to set and recommend pricing, and responds to any fluctuations in such data by making automatic changes to prices.
Broadly, such algorithms generate two modes of pricing:
- Dynamic (or surge) pricing: this software varies prices in real time based on changes to supply and demand. This practice is used widely in sectors such as air travel, hotel bookings, and e-commerce.
- Personalised pricing: whilst dynamic pricing pulls data from market demand, personalised pricing algorithms tailor prices to individual customers by relying on personal data, browsing history, and previous purchases.
Risks to Competition
The fundamental issue with algorithmic pricing derives from Chapter I of the Competition Act 1998 (the “Competition Act”), which prohibits any agreement or concerted practice between undertakings that has the object or effect of distorting competition in the UK. This prohibition captures the exchange of commercially sensitive information between actual or potential competitors, as a key principle of competition law is that each business should independently form its own strategy without the influence of other firms; if undertakings receive information enabling them to predict how others will act on the market, they may alter their behaviour to align with other businesses and consequently reduce customer choice.
The CMA has published substantial guidance on the risks of algorithmic pricing to competition, identifying the potential for such tools to harm consumers through rapid fluctuations in price, and even to facilitate unlawful coordination between rival undertakings in breach of the Chapter I prohibition. Scrutiny from the authority has focused particularly on this algorithmic collusion, with a recent blog post titled ‘AI and collusion: frontiers, opportunities and challenges’ outlining four scenarios in which this issue may arise:
- ‘Classic’ collusion: competing businesses may explicitly agree to collude, and then use algorithmic tools to enforce such agreements;
- ‘Hub-and-spoke’ collusion: businesses may use algorithms or data hubs as a platform to indirectly exchange commercially sensitive information, even delegating pricing decisions to these hubs in certain instances;
- ‘Predictable agent’ collusion: algorithms that are designed to react predictably to market events may develop to follow price leadership, resulting in collusion without the need for human communication or any explicit agreement; and
- ‘Autonomous’ collusion: pricing algorithms have the capability to independently learn to collude through agentic AI systems, creating the potential for such tools to coordinate strategies without any human input.
Recent Investigations
On 24 February 2026, the CMA launched an investigation into the suspected sharing of competitively sensitive information between rival hotel chains using CoStar’s data analytics tool ‘STR’. The investigation, focusing on CoStar, Hilton, IHG, and Marriott, serves as the CMA’s first material action against algorithmic-based breaches of the Competition Act in the hotel sector.
The authority has also revealed increased investment in its technological capabilities, including the use of AI and agentic systems to screen for algorithmic collusion ‘at an unprecedented pace and scale’, and an updated leniency policy that makes leniency available for the exchange of commercially sensitive information via a shared algorithm and offers up to £250,000 to anyone who informs the CMA of cartel activity (including algorithmic pricing). These developments, combined with the authority’s ability to test and trial algorithms directly under the Digital Markets, Competition and Consumers Act 2024, send a clear warning to businesses that algorithmic pricing is placed as a key priority for the CMA.
Outside of the UK, pricing algorithms have been subject to increased attention from regulators globally. The US Department of Justice set the stage for enforcement action in August 2024, launching a lawsuit against RealPage on the basis that the revenue management software provider had facilitated price fixing; by commingling pricing data and using competitively sensitive information to set rental rates, the provider enabled competing property management companies to align on pricing. Whilst no details have been publicly disclosed yet, a senior official from the European Commission confirmed in July 2025 that several investigations are currently being conducted into algorithmic pricing.
Key Takeaway
Whilst it remains to be seen whether ongoing investigations across Europe and the UK will conclude with any penalty action, these recent probes signal a clear shift in global regulators’ approach; algorithmic pricing is no longer solely an area of academic interest for competition authorities, but rather an active enforcement priority.
Atlantis O’Neill is a Trainee Solicitor at Dentons and a member of the SCL Trainee Group
AI in Law: From Experiment to Everyday Reality
Not long ago, talking about artificial intelligence in a law firm would have earned you a polite smile and a quiet eye roll. That feels like a lifetime ago now. In 2025, generative AI stopped being a background experiment and became something legal teams actually depend on, day in, day out. Nobody really predicted how fast it would happen.
The Tools Everyone Is Talking About
Two platforms keep coming up in conversation: Harvey and Legora. They are doing quite different things, and both are worth paying attention to.
Harvey has become the go-to tool for large firms handling complex, document-heavy work. Its Vault feature can analyse up to 100,000 documents in a single collaborative workspace, with customisable workflows for tasks like due diligence and drafting. By the end of 2025, Harvey was valued at $8 billion, which, whatever you think about tech valuations, says something about where the market sees this heading. Daily active usage had grown by 81% since 2023.
Legora takes a different approach. Where Harvey leans into legal reasoning and analysis, Legora is built around execution at scale – helping teams run contract review as a structured, repeatable process across thousands of documents. It has a tabular review feature that organises extracted data in a way that will feel familiar to anyone who has ever lived in a spreadsheet, and a Word add-in that brings rule-based playbooks into the drafting environment. It is less flashy than Harvey, but for teams drowning in volume work, that is not the point. Both platforms take data security seriously, which matters when you are handling sensitive client information.
Contracts: Where AI Has Made the Biggest Dent
If there is one area where AI has genuinely changed the day-to-day, it is contracts. That is not surprising. Contracts are structured, repetitive, and time-consuming in ways that play to AI’s strengths. Tools like Spellbook, which embeds directly into Microsoft Word, now offer real-time clause suggestions, risk flags, and a “Compare to Market” feature that benchmarks terms against thousands of comparable agreements. For junior lawyers who once spent days grinding through first-draft reviews, the difference is hard to overstate.
Research: Impressive, but Watch the Hype
Legal research has also been transformed, though the evidence here is still catching up with the hype. The Vals Legal AI Report, the first independent benchmark study of its kind, tested Harvey, CoCounsel, Vincent AI, and others across six research task categories. Harvey led in five of the six, achieving 94.8% accuracy on document Q&A. The speed gains were arguably even more impressive: AI tools completed tasks between six and eighty times faster than lawyers working manually. Those are the kinds of numbers that make you sit up.
Clio’s $1 billion acquisition of vLex -the largest M&A deal in legal tech history for a privately held company – underlined just how valuable AI-powered research has become. The combined platform now serves 400,000 legal professionals and sits at a $5 billion valuation. Whether that kind of consolidation is good for the market in the long run is another question, but it certainly concentrates minds.
Meanwhile, LexisNexis and Thomson Reuters have both doubled down on what they call agentic AI -systems that do not just answer questions but plan, execute, and self-evaluate multi-step tasks. LexisNexis has set itself a target of automating 15 to 20% of lawyer tasks through its Protégé platform by 2028, which is ambitious, to put it mildly. Thomson Reuters positioned its CoCounsel Legal platform as “the most comprehensive AI solution for legal professionals to date,” though vendors are rarely shy about that sort of claim.
Everyone Is Using It. Almost Nobody Is Governing It.
The numbers on adoption are striking. Nearly seven in ten legal professionals now use generative AI in their work, a figure that more than doubled in a single year. Among regular users, 38% save between one and five hours per week, and 14% save six to ten hours weekly. To put that in perspective, that is the equivalent of gaining back half a working day or more -every week.
But the governance picture is, frankly, worrying. More than half of firms -54% -have provided no training on responsible AI use and have no plans to do so. Only 9% have a written, actively enforced AI policy. Courts have already sanctioned practitioners for submitting AI-generated content that included fabricated case citations. These are not hypothetical risks. The profession’s ethical duties around competence, confidentiality, and candour do not pause just because the technology is exciting.
So Where Does That Leave Us?
Legal AI is not a future trend anymore. It is here, the productivity gains are real, and the competitive pressure is only going in one direction. But the harder problem is not the technology -it is the governance, the training, and the cultural shift needed to use it properly. Firms that get that right will do well. The ones that do not are storing up problems for themselves.
Savvas Skordellis is a Trainee Solicitor at Deloitte and a member of the SCL Trainee Group
Deepfakes, Disclosures and Deadlines: Engineering AI Transparency Before 2 August 2026
On 17 December 2025, the European Commission published a first draft Code of Practice on Transparency of AI-Generated Content, intended to support compliance with the content-related elements of the EU AI Act’s Article 50 transparency obligations. The Commission’s timetable is compressed: feedback on the first draft closed on 23 January 2026, a second draft is due mid-March 2026, and the Code is expected to be finalised by June 2026. The underlying transparency rules become applicable on 2 August 2026.
The draft Code’s core message is practical: transparency cannot be treated as a one-off label added at the point of generation. It needs to hold up when content is shared, reformatted, edited, compressed, or re-uploaded. In other words, transparency needs to survive the real-world content lifecycle—because approaches that only work within your product are likely to break once outputs leave your environment.
What the draft Code is trying to standardise
The Commission describes the draft Code as having two sections, with different responsibilities across the value chain:
- Marking and detecting AI content (providers): measures aimed at making outputs machine-readable and detectable as artificially generated or manipulated.
- Labelling deepfakes and certain public-interest text (deployers): measures aimed at ensuring people are clearly informed when they are seeing (i) a deepfake, or (ii) certain AI-generated or manipulated text published on matters of public interest.
In practice, many organisations will sit in both categories – offering tools that generate or manipulate content while also publishing that content or enabling others to publish it. That is why the split matters most at the hand-off: a company may implement technical marking, but still deliver a user experience where disclosures are inconsistent, easy to miss, or lost downstream.
The uncomfortable reality: “perfect labelling” is not the standard
The draft Code implicitly recognises a simple constraint: you cannot “solve” transparency with a single technology (such as a watermark) and assume the job is done. In live distribution environments:
- Watermarks can be degraded or removed through cropping, compression, or re-encoding.
- Metadata can be stripped by platform pipelines, screenshots, or copy/paste.
- Detection can be undermined through deliberate manipulation, remixing, or model-mixing.
The compliance question is therefore less “is every output always detectable?” and more: have you adopted reasonable, repeatable measures aligned to your key use cases, and can you show you designed for predictable failure modes?
What “good” looks like in practice
Stripped of legal framing, the draft Code points towards four characteristics of a credible transparency posture:
- Layering: more than one transparency mechanism, because any single method will fail in some flows.
- Consistency: aligned signals across UI, exports, and APIs so downstream systems can preserve transparency.
- Prominence: disclosures that are hard to miss in high-risk contexts (deepfakes; public-interest communications).
- Evidence: documentation of design choices, testing, and what happens when transparency mechanisms break.
A practical way to act without building a bureaucracy
Most organisations will not need an expansive, multi-workstream compliance programme to start making progress. A sensible initial approach is to focus on three deliverables, and then iterate as the Code develops and products mature:
- A scope map: identify where AI-generated or manipulated content is created, edited, exported, and published – paying particular attention to the transformations most likely to strip or degrade transparency signals.
- A transparency standard: a short internal specification that defines (a) what gets marked, (b) what requires disclosure, (c) how and where disclosures appear (in-product and on export), and (d) what transparency signals are passed through APIs to downstream users.
- A resilience test: a lightweight test suite that checks whether your transparency signals survive common formats and transformations (e.g., compression, cropping, re-encoding, reposting). Where they do not, you have a structured basis for adding a second layer (for example, pairing metadata with visible disclosure in relevant contexts).
Takeaway
Treat the draft Code as an early indication of what regulators are likely to view as a credible transparency posture by 2 August 2026: not simply “we added a label”, but “we engineered transparency across the content lifecycle.” If your approach does not survive export and redistribution, it is unlikely to survive scrutiny.
Safwan Akbar is a Trainee Solicitor at Morrison Foerster and a member of the SCL Trainee Group
Levelling the Playing Field – GenAI and Gaming
Game developers come in all shapes and sizes. From first party developers (Nintendo, Sony, Xbox) and third-party heavyweights (EA or Ubisoft) to independent or ‘indie’ developers shipping from a single laptop. Traditionally, market influence has scaled with budget and headcount. However, the advent of AI, and in particular, generative AI, is levelling the playing field. Solo developers are 3x more likely to use AI art tools than larger teams.
But in the UK and across Europe, the promise of GenAI will, in part, depend on how three legal fronts settle:
- data protection concerns relating to AI training;
- copyright issues at both dataset and model layers; and
- application of the EU AI Act that will increasingly shape the AI tools that game teams rely on.
This blog update looks at the legal considerations in play and how the regulatory landscape may unfold.
Data Protection
Europe’s most concrete signal to date has come from Germany. In May 2025 the Higher Regional Court of Cologne refused to block Meta’s plan to use publicly accessible Facebook/Instagram posts to train AI models. The court accepted legitimateinterests under the GDPR where users were informed and could object. The court also saw no breach of the Digital Markets Act on the facts. Whilst this was an interim‑relief decision, it shows that legitimate interest‑based training on public data, with transparency and objections, can pass muster.
The European Data Protection Board’s late‑2024 opinion pushed in the same direction but set a high bar. Controllers must show strict necessity and pass a balancing test, and DPAs will scrutinise claims that models are “anonymous.” In the UK, the ICO adopts a comparable stance. Legitimate interests remains the only practical basis for current scraping, but the ICO expects developers to justify why scraping is necessary over licensed sources and to reduce “invisible” processing through clear transparency and workable opt‑outs.
Copyright
Despite the opportunity AI presents, 42% of indie developers cite “copyright infringement” as their primary hesitation when using AI. Do they have cause for concern?
The UK’s headline case, Getty Images v Stability AI didn’t settle whether English law treats model training on copyright works as infringement, but it did indicate how risk should be managed. It held that an AI model can count as an “article” for UK secondary‑infringement purposes, however, as established in the case, they aren’t necessarily articles. The court also found only limited trademark infringement tied to outputs that reproduced Getty’s marks/watermarks.
The immediate read‑across is that tight filtering, deduplication and watermark/brand‑suppression on outputs materially reduce exposure to IP claims, and good provenance controls help demonstrate mitigation. At the same time, because the court didn’t rule on UK training liability, that question remains open for a better‑framed case focused on UK‑based acts of copying.
EU AI Act
On the game surface itself, Article 50 of the EU AI Act will require clear signalling when players interact with an AI system. The Commission’s second draft Transparency Code takes a multilayer approach. It covers metadata, imperceptible watermarking, and, where necessary, fingerprinting/logging, plus detectors and visible labels for deepfakes and AI‑generated text used to inform the public. The UK has not enacted an EU‑style horizontal AI law and continues to take a principles‑based, regulator‑led approach, with practical playbooks rather than prescriptive duties.
The implementation of the Act remains to be seen, but the signalling requirements may influence how different studios approach AI. Some indie teams may find the additional transparency steps resource‑intensive. Equally, implementing these measures could reassure players and rightsholders by providing greater clarity around synthetic content. Overall, the impact is likely to vary depending on the tools developers use, the scale of their workflows, and how the Commission refines guidance in practice.
Next steps
For UK and EU developers alike, the direction of travel seems consistent. Teams relying on AI, whether for assets, prototyping or localisation, should track how data protection, copyright, and AI‑specific regimes converge. A practical starting point includes;
- auditing the source of datasets and tools;
- implementing transparency measures that can scale as rules become clear; and
- engaging early with vendors to understand how they intend to meet EU AI Act duties.
As the regulatory landscape settles over the next 18–24 months, studios that build compliance into their pipelines now will be better placed to harness GenAI’s advantages without inheriting avoidable risk.
Avi Marcus is a Trainee Solicitor at DLA Piper and a member of the SCL Trainee Group
Blue Links, not Black Boxes – Competition Investigation opened into Google’s AI Summaries
If you type a search query into Google (or indeed, most major search engines), you will see at the top of the page an “AI Overview” box, with the usual blue links to individual websites (sometimes called “organic results”) following after.
Screenshot: AI Overview’s response to search query “holiday destinations in February”
On 9 December 2025, the European Commission opened a formal antitrust investigation to assess, among others, whether Google’s use of web publishers’ content to provide AI summaries could constitute an abuse of dominance under Article 102 TFEU. In particular, the Commission was concerned that Google may have been doing so (1) without providing appropriate compensation to publishers, and (2) without offering the possibility to refuse such use of their content without losing access to Google Search.
Anticompetitive effects / pro-competitive justifications
Several studies point to likely anticompetitive effects of this behaviour. For example, one recent study by the Pew Research Centre suggested that people only clicked a link once in every 100 searches when there was an AI summary at the top of the page. A separate study by Bain & Company notes that 80% of consumers now rely on AI-written results, with a corresponding reduction in organic web traffic by 15-25%. Publishers like the Daily Mail have also claimed that the number of people who visit its website from Google Search results have fell by about 50% since Google introduced its AI overview feature.
This represents a concern for web publishers or content creators who may rely on Google Search for a substantial amount of web traffic (and hence ad revenue, user visibility etc). This is especially concerning, it is argued, if this comes as a result of the use of publishers’ / creators’ works to train the AI summary model, without options to opt-out without losing access to Google Search.
Pro-competitive justifications will likely be offered in response – e.g. that AI summaries improve the functionality or value of Google Search, that dong so is indeed pro-competitive as it results in better consumer outcomes and welfare, or perhaps point to the fact that Google’s AI summaries attempt to link the user to original source webpages. It would also need to be shown that any possible justifications could not have been achieved through less restrictive alternative means.
Big tech antitrust cases, other copyright issues
We are unlikely to see the results of the Commission’s investigation in the immediate future, and any possible litigation, subsequent appeals, and follow-on claims will likely take many years to resolve. In the meantime however, it is worth noting that big tech players are certainly familiar with competition law investigations into such conduct – for a nostalgic example see the Commission’s decision to fine Microsoft for tying the Internet Explorer browser to Windows. A proper assessment in each case of what is likely to amount to anti-competitive behaviour will involve fact-intensive exercises, requiring very significant amounts of evidence, which can only be properly evaluated in court if the Commission decides to issue a case following the investigation.
Readers following the AI regulation space will also notice that Google’s behaviour raises parallel but distinct issues with respect to copyright law. Similar arguments regarding compensation and opt-out provisions have been run with respect to unauthorised AI training and copyright law (see e.g. the UK Government’s ongoing AI / IP consultation, or recent High Court judgment in Getty Images v Stability AI). A very substantive amount of commentary surrounding these issue has been written – and as regulators catch up with enforcement measures we get ever-closer to seeing how the hammer will fall.
Solomon Chann is a Trainee Solicitor at Bristows and a member of the SCL Trainee Group
