Cyprus launched a public consultation on its National Strategy for Artificial Intelligence.
Australia announced it will shift from voluntary to mandatory rules for artificial intelligence systems.
Tempus AI agreed to buy cancer testing firm Personalis in a deal worth over one billion dollars.
Chinese lab Moonshot AI launched the world's largest open weight model with nearly three trillion parameters.
Researchers warned that techniques used to make AI safe could be turned into tools for government censorship.
Policy & Regulation
Cyprus launches public consultation on National AI Strategy
As announced by the Deputy Ministry of Research on 22 July, Cyprus has launched a public consultation on its proposed National Strategy for Artificial Intelligence, with submissions accepted until 31 August. Prepared by the National AI Taskforce, the strategy aims for Cyprus to become a trusted AI hub in the Eastern Mediterranean, a reliable European jurisdiction for AI-enabled services and a bridge between the European Union and neighbouring regions by 2032. Its priorities include responsible AI adoption, productivity, public-sector transformation, skills development, secure data infrastructure and stronger governance, ethics and accountability. Businesses and other stakeholders should consider participating in the consultation and monitor the final strategy, which is expected to influence future AI policy, investment, infrastructure and public-sector initiatives in Cyprus.
Australia Transitions to Mandatory AI Governance Framework
Australia is moving from voluntary ethics principles to a mandatory legal framework for artificial intelligence. Prime Minister Anthony Albanese announced the shift on 15 July following a national consultation. The government concluded that AI requires purpose-built regulation rather than the previous reliance on voluntary safety standards. Plans are now underway to establish a dedicated Office of AI and introduce legislation early next year. Businesses operating in Australia should prepare for binding obligations as the government seeks to manage AI risks more effectively through formal oversight.
Greece enacted Law 5321/2026 on 20 July to establish its national framework for enforcing the EU AI Act. The legislation names the Hellenic Data Protection Authority as the primary market supervisor and national contact point. It also creates a regulatory sandbox for testing new AI tools and introduces transparency measures for systems used in the public sector. This law provides a structure for complaints, investigations, and enforcement within the country. Companies deploying AI in Greece must now identify their specific regulators and prepare for more active supervision.
European Commission Issues AI Transparency Guidelines
The European Commission released guidelines on 20 July regarding transparency rules under the EU AI Act. These rules take effect on 2 August 2026. The guidelines help people identify when they are interacting with AI or viewing content generated by it. Providers and users must disclose the use of deepfakes, biometric systems, and emotion recognition tools. This measure aims to reduce deception and manipulation in matters of public interest. Organisations should review these requirements to ensure they properly label synthetic media before the enforcement deadline.
Tempus AI announced on 20 July that it will acquire cancer-testing company Personalis in an all-stock deal valued at approximately $1.5 billion. The acquisition aims to integrate advanced molecular testing into the Tempus AI-driven precision oncology platform. The two companies have collaborated since late 2023, and Tempus is already a major shareholder. This move reflects a broader trend of consolidation in the healthcare diagnostics industry. The transaction is expected to close by early 2027, though investor caution remains regarding the integration of the two diagnostic platforms.
Semiconductor startup Eliyan raised $145 million in Series C funding on 29 July. The round values the company at $1 billion and was led by Seligman Ventures with support from major firms like Meta and Intel. Eliyan develops technology to speed up data transfer between chips in data centres. This addresses a critical bottleneck as AI clusters grow larger. The funding will help Eliyan expand into optical interconnects to reduce power use. Investors remain eager to fund AI infrastructure hardware despite broader concerns about overspending on data centres.
Chinese lab Moonshot AI released Kimi K3 in July 2026. With 2.8 trillion parameters, it is the largest open weight model released to date. The model uses a mixture-of-experts design to keep running costs manageable by only activating 16 experts per query. While it trails leading proprietary systems, it outperforms other open models on independent benchmarks. This release suggests the performance gap between open and closed source AI is closing rapidly. Developers can now download the full weights to build their own applications.
Award Winning Paper Warns of AI Safety Tool Misuse
Researchers Sarah Ball and Phil Hackemann won a top award at ICML 2026 for their paper on AI safety tools. They argue that alignment techniques meant to keep AI safe could be repurposed for digital censorship. If an authority gains control over these tools, they could use them to suppress legitimate speech. The authors did not name specific governments but warned the risk applies to any entity with control over AI systems. They urge the research community to build safeguards into safety tools now to prevent future misuse.
