Professional profile
About Tian
Generative AI, Algorithms and Intellectual Property in Guangzhou
Tian Wu advises innovators, technology companies and content businesses on patent protection, AI-related intellectual property and creative-industry compliance from Guangzhou’s Tianhe technology corridor—one of southern China’s most active centres of software, internet services and advanced manufacturing.
As a lawyer at Guangdong Zhihang (Guangzhou) Law Firm, Tian Wu combines long courtroom and advisory experience across civil, commercial and related practice with a formal qualification path as a Chinese Patent Agent. He has practiced since 2000—more than two decades in the profession—and works in English and Mandarin with Guangdong Bar Association (Guangzhou) credentials on file. Clients come to him when generative tools, training data and model outputs collide with patent, copyright and trade-secret questions that standard product counsel have not yet operationalised.
AI files are rarely “one statute.” They combine inventorship and patentability questions, ownership of prompts and outputs, dataset licensing, open-source contamination risk, and platform terms that shift liability to enterprise users. Mr. Wu’s method is to pin the commercial use case first—what is sold, what is automated, what data is ingested—then choose IP and contract tools that match that reality.
Patent Strategy for AI-Related Inventions
He advises on claim strategies that survive abstract-idea style objections, coordination between software architecture disclosures and trade-secret retention, and timing of filings against product launch calendars in the Greater Bay Area. For companies collaborating with universities or vendors in Guangzhou and Shenzhen, joint-development agreements must allocate improvement ownership before models are fine-tuned on shared data.
Invalidation and freedom-to-operate reads are treated as business decisions: which claims block a competitor, which licences are cheaper than redesign, and which markets justify the cost of multi-jurisdiction filing.
Copyright, Content and Algorithm Governance
Generative content businesses need clear rules on training sources, output ownership for enterprise customers, and takedown response when third parties claim infringement. Mr. Wu drafts customer terms and internal playbooks that product and trust-and-safety teams can apply. Where algorithm recommendations or automated decision features affect consumers, he coordinates with privacy counsel so IP documents do not contradict personal-information notices.
- AI-related patent prosecution and portfolio counselling
- Copyright and content ownership for generative systems
- Trade-secret protocols around models, weights and datasets
- Commercial contracting for AI features sold into enterprises
Local Practice Notes and Engagement
Guangzhou matters often involve manufacturing clients embedding AI into equipment and consumer apps distributed nationwide. Enforcement strategy must consider where servers, defendants and evidence sit. Mr. Wu confirms scope in writing and provides bilingual summaries for overseas general counsel. Contact this profile with a product description, data sources used for training or fine-tuning, and whether you face an active infringement claim or a launch-blocking FTO question.
Manufacturing AI, Customer Contracts and Bay Area Enforcement
Guangzhou and Pearl River Delta manufacturers embedding vision models and predictive maintenance AI need IP allocations with system integrators. Mr. Wu drafts statements of work that separate pre-existing tools, project-specific models and customer data. Without that split, both sides claim the same weights and datasets when the relationship ends.
Enterprise buyers increasingly demand audit rights on training data legality and security controls. He helps vendors respond with evidence packages that are honest about residual risk rather than absolute warranties no engineering team can support. Overpromising in security and IP warranties is a litigation seed he tries to remove at contracting.
Enforcement against copycat devices sold online benefits from a combined patent-copyright-unfair competition approach. He selects the lead theory based on evidence strength and speed to injunction, not based on which legal label sounds most modern. Generative AI buzzwords do not impress tribunals without a clean factual record.
For content platforms using generative features, he aligns user terms, creator revenue shares and notice-and-takedown workflows so that IP operations and product growth are not at war. Training sessions for product counsel focus on decision gates before a model is fine-tuned on a new corpus.
Standards, Open Source and Litigation Readiness
Mr. Wu monitors how standards bodies and industry codes discuss AI risk, translating them into contract schedules enterprise buyers recognise. He helps clients avoid certification theatre that claims alignment without evidence trails.
Open-source licence conflicts in model toolchains are handled with inventory and cure plans. Litigation readiness means keeping development logs, dataset licences and model cards in a repository counsel can access under privilege-aware processes when a subpoena or lawsuit arrives.
He closes engagements with a prioritised risk register: must-fix before launch, fix within one quarter, accept with monitoring. That register is the artefact boards understand, and it is more useful than a narrative essay about the future of AI law.
Capability
