
5 Things to Know About AI This Week
A quick scan of what happened in AI this week.
1. China launches a global AI governance body with 29 member countries
At the World AI Conference in Shanghai on July 17, Xi Jinping announced the World Artificial Intelligence Cooperation Organization (WAICO), an intergovernmental body headquartered in Shanghai with 29 founding members including Russia, Brazil, Indonesia, and South Africa. Xi pitched it as an alternative to US-led AI frameworks, criticized American export controls, and pledged 5,000 AI training slots for developing countries over five years. The US-led Pax Silica initiative currently counts 35 signatories.
Real estate angle: The practical takeaway is that AI regulation is splitting into competing blocs, and the tools you use will increasingly carry a flag. If a vendor's product runs on models hosted or governed abroad, ask where transaction data actually lives and which jurisdiction's rules apply to it. Wire instructions, buyer financials, and NPI do not belong in systems you cannot trace.
Source: Fortune
2. Moonshot AI releases Kimi K3, the largest open-weight model yet
China's Moonshot AI launched Kimi K3 on July 16, a mixture-of-experts model with roughly 2.8 trillion total parameters, a 1-million-token context window, and native vision. API pricing is $3 per million input tokens and $15 per million output, undercutting most domestic models. Full open weights are scheduled for release by July 27 under a modified MIT license.
Real estate angle: Cheap, capable, openly licensed models mean the software vendors serving our industry can swap the AI engine under the hood without telling anyone. Your CRM, your listing description generator, or your document summarizer may be running a different model next month than it ran last month. Contracts with technology vendors should address which models process your data and require notice when that changes. If your vendor cannot answer the question, that tells you something too.
Source: Constellation Research
3. Meta kills its Instagram AI image tool after three days of backlash
Meta launched a feature on July 7 that let any user generate AI images by tagging public Instagram accounts as visual references, on by default for all public accounts over 18. Users had to find the setting and opt out to keep their photos from feeding other people's image generations. After protests from users and a sharp statement from SAG-AFTRA, Meta pulled the feature on July 10.
Real estate angle: This is important if you run a public Instagram account marketing yourself and your listings. Social media platform's terms can change overnight and they typically default to taking date, and not asking for permission. Audit what you post publicly. There is also a liability edge: if someone generates a fake endorsement or a doctored listing photo from your public content, you want a record that you never consented. Screenshots and timestamps are cheap insurance.
Source: Variety
4. Hawaii and Illinois sign new AI laws, and listing photos are next
Hawaii Gov. Josh Green signed two AI bills on July 13: SB 3001, a chatbot safety law requiring operator disclosures, and HB 2137, a deepfake law that prohibits harmful digital imitations and requires disclosure of synthetic performers in advertising. Illinois Gov. Pritzker signed two more the same week. Meanwhile, California's AB 2025, which would require disclosure when AI digitally alters promotional materials for the sale of real property, has an appropriations hearing set for August 3, and New Jersey's A 4728 targets deceptive AI in dwelling advertising.
Real estate angle: Virtually staged photos, AI-brightened rooms, and generated twilight shots are standard practice in listing marketing, and legislatures are moving to require disclosure. Brokers should get ahead of this now: adopt a written policy on AI-altered listing media and disclose the alteration in the listing itself. An undisclosed AI-staged photo is a misrepresentation claim just waiting for a plaintiff. Disclosure costs one line of text.
Source: Transparency Coalition
5. The AI money keeps flowing: over $1 billion in a single day
Venture funding into AI startups showed no summer slowdown. On July 15 alone, roughly $1.02 billion was announced across eight AI startups spanning healthcare, infrastructure, identity, voice, and enterprise automation. The day before, AI video company AIsphere closed a $439M Series C led by Alibaba, and Chai Discovery raised $400M from Index, Kleiner Perkins, and Sequoia. Both Anthropic and OpenAI moved toward IPO filings earlier this month.
Real estate angle: All of this funding means a flood of new AI tools will keep landing in your inbox, each promising to transform your business. Some will be gone in 18 months. Before you move client data or core workflows into a startup's product, ask the unglamorous questions: who owns the data if they fold, is there a way to export my data, and does your E&O carrier have a position on AI-assisted work product. Adopt tools on a trial basis and have an exit plan.
Source: Tech Startups
The theme this week is control: who controls AI governance, who controls your likeness (you or the platform), and who controls what an AI-altered listing photo must say about itself (increasingly, state legislatures). The technology is racing ahead while the rules get written in real time, state by state. For real estate professionals, the safest posture has not changed: disclose what AI touched, know where your data goes, and keep a paper trail. The firms that treat those as habits now will not be scrambling when their state adopts similar disclosure laws.
- Jason