Quick Brief
Two developments defined the technology policy landscape in late April 2026. In U.S. court proceedings tied to his lawsuit over OpenAI’s direction, Elon Musk disclosed on 30 April that xAI’s Grok model was trained using distillations of OpenAI models, placing competitive and legal scrutiny on AI training practices, according to TechCrunch. In Europe, the European Commission published its first review of the Digital Markets Act (DMA) on 28 April and, one day later, issued a preliminary finding that Meta failed to prevent under‑13s from using Instagram and Facebook under the Digital Services Act (DSA), per the European Commission.
Musk’s testimony sharpened questions about how frontier AI developers compete, where the line sits between lawful model use and prohibited copying, and what "distillation" implies for IP and contract boundaries. In the EU, the juxtaposition of a DMA progress review and a DSA enforcement step underscored how market-structure and content-safety regimes are moving in tandem, affecting gatekeeper platforms and social networks in distinct but complementary ways Digital Markets Act review DSA Meta underage users.
xAI’s training disclosures surface in OpenAI courtroom fight
On 27 April 2026, Elon Musk filed a lawsuit challenging OpenAI’s direction, governance, and control, setting the stage for testimony focused on the company’s evolution and the use of powerful AI models, according to the AP. Opening statements on 28 April emphasized control and governance themes, as reported by NPR. By 29 April, Musk testified that OpenAI’s direction had departed from its original purpose, per NPR, and continued under questioning connected to his bid to legally dismantle OpenAI, according to Axios.
The inflection point came on 30 April, when Musk stated under oath that xAI trained Grok using distillations of OpenAI’s models, as reported by TechCrunch. The disclosure centers debate on whether distillation from a proprietary model—potentially learned via outputs or intermediate behaviors—raises intellectual property or contractual issues when the source model is not open. WIRED separately characterized Musk’s statement as an apparent admission that OpenAI models informed Grok’s training, a framing that could influence how rivals, partners, and courts interpret the scope of permissible training inputs.
Witness credibility also entered the frame. On 30 April, reporting highlighted inconsistencies in Musk’s testimony about xAI’s internal "safety cards" and his temperament on the stand, per Ars Technica. Those characterizations contrast with Musk’s own narrative about OpenAI’s mission drift and governance, which he advanced on 29 April, according to NPR. The record on 29 April also included color about personal history between the principals, with courtroom remarks about an old friendship, as described by TechCrunch.
From an industry standpoint, the distillation disclosure (30 April) matters because it touches how fast followers may attempt to match frontier capabilities without full-stack original training—and how boundaries are drawn between fair development practices and misuse of proprietary systems. The testimony itself "raised legal and competitive questions about AI model distillation" on 30 April, according to TechCrunch. While the legal process will determine the weight of these statements, companies building with or alongside closed models will closely watch how courts and counterparties interpret such training techniques.
Key developments
- (2026-04-27) Musk filed a lawsuit over OpenAI’s direction, governance, and control, according to the AP.
- (2026-04-28) Opening statements focused on control and governance at OpenAI, per NPR.
- (2026-04-29) Musk testified that OpenAI’s direction had departed from its original purpose, according to NPR.
- (2026-04-30) Musk stated that xAI trained Grok using distillations of OpenAI models, per TechCrunch.
- (2026-04-30) Reporting flagged inconsistencies in Musk’s testimony about xAI’s "safety cards," noted by Ars Technica.
EU tightens platform oversight: DMA progress and a DSA step on Meta
On 28 April 2026, the European Commission published the first formal review of the Digital Markets Act, setting out early signs of how gatekeeper obligations are reshaping software and device ecosystems, according to the European Commission. In that review dated 28 April, the Commission said alternative app stores have launched as mobile operating systems opened to third‑party distribution, users can choose non‑default search engines and web browsers, and people can transfer their data when switching services and devices; the document also described enhanced interoperability for manufacturers of connected devices, per the European Commission. Separately on 28 April, the Commission highlighted the DMA as "fit for purpose" with positive competitive impact, as publicized on its digital‑strategy site, according to the European Commission.
On 29 April, the Commission announced a preliminary finding that Meta breached the DSA by failing to prevent minors under 13 from accessing Facebook and Instagram, marking a child-safety enforcement step under the EU’s online content regime, according to the European Commission. Also on 29 April, a separate account stated the EU alleged Meta violated digital rules requiring platforms to protect minors, per NPR. Because the 29 April communication is a preliminary finding, the process suggests a path to remedial measures after Meta’s response; the Commission’s framing underscores that the step precedes any final decision.
Taken together, the 28–29 April milestones highlight the EU’s dual-track approach: market structure and competition via the DMA alongside content moderation and safety via the DSA. The DMA review’s early interoperability and choice signals (28 April) align with a policy push to lower switching costs and encourage third‑party app distribution. The DSA preliminary finding (29 April) emphasizes child protection obligations and raises operational questions for large platforms about age verification, policy design, and enforcement at scale. Media and institutional accounts differ in emphasis: the Commission’s DMA material (28 April) highlights systemic changes like alternative app stores, while the 29 April DSA step homes in on specific compliance gaps for a single platform DMA review highlights DSA Meta breach.
Key developments
- (2026-04-28) The European Commission published the first review of the Digital Markets Act, according to the European Commission.
- (2026-04-28) The DMA review said alternative app stores have launched as operating systems opened to third‑party stores and that users can choose non‑default browsers and search engines, per the European Commission.
- (2026-04-28) The review also described data portability improvements for users and enhanced interoperability for connected‑device manufacturers, according to the European Commission.
- (2026-04-28) The Commission highlighted that the DMA remains fit for purpose with positive impact on competition and innovation, per the European Commission.
- (2026-04-29) The Commission preliminarily found Meta in breach of the DSA for failing to prevent under‑13s on Instagram and Facebook, according to the European Commission.
- (2026-04-29) A separate account stated the EU alleged Meta violated digital rules requiring protection of minors, per NPR.
Conclusion
Disclosures and enforcement steps late in April 2026 tightened focus on two fault lines in tech. In the United States, Musk’s 30 April testimony about distilling OpenAI models to train Grok—set against a lawsuit filed on 27 April and governance‑oriented opening statements on 28 April—pushed distillation and model provenance from technical practice into legal argument, with competitive implications, as reported by TechCrunch, NPR, and the AP. In the European Union, the 28 April DMA review framed early signals of increased user choice and interoperability, while the 29 April DSA preliminary finding on Meta underscored intensifying scrutiny of child-safety compliance, per the European Commission and NPR. For industry leaders, the through-line is clear: governance choices around how models are trained and how platforms are operated are becoming testable in courtrooms and regulatory proceedings, with outcomes that will shape competitive strategies and compliance playbooks alike.
