5 Things to Know About AI This Week
A quick scan of what happened in AI this week.
1. Anthropic files confidentially for an IPO
On June 1, Anthropic submitted a draft Form S-1 to the SEC, officially kicking off the process to go public. The company's revenue run rate hit $47 billion in May, up from roughly $10 billion the prior year. Its last private funding round valued the company at $965 billion, topping OpenAI's $852 billion valuation from March. OpenAI is expected to file its own S-1 in the coming weeks.
Real estate angle: Two of the biggest AI companies in the world are heading to public markets, and that means disclosure requirements, earnings pressure, and shareholder scrutiny that private startups never face. For real estate professionals building on these platforms, more financial transparency is a good thing. It also means pricing changes and new product tiers are likely as each company chases the revenue it needs to justify its valuation. Check what your AI contracts say about price increases before you find out the hard way.
Source: CNBC
2. Microsoft launches its first in-house AI model family
At Microsoft Build (June 2-3), the company unveiled seven proprietary models under the MAI brand. The flagship, MAI-Thinking-1, is a reasoning model that Microsoft says matches Anthropic's Opus 4.6 on coding benchmarks and outperforms Sonnet 4.6 in blind evaluations. The lineup also includes dedicated models for image generation, transcription, and voice synthesis. Microsoft also said it is repositioning Windows as a first-class environment for autonomous AI agents.
Real estate angle: Microsoft has resold OpenAI's models for years, and building its own changes the relationship. If you run Copilot in Microsoft 365 for document drafting, email, or scheduling, expect faster feature changes and tighter integration. It also means the model behind your Copilot may change without much notice. Ask which model is handling your data.
Source: Microsoft Build 2026
3. New York bill would make chatbot operators liable for unauthorized professional advice
New York Senate Bill S7263 reached the Senate floor calendar in late February and moved to a third reading in March. The bill would bar operators of AI chatbots from giving substantive responses that would count as the unauthorized practice of a licensed profession, including law, if a human gave them. It creates a private right of action for actual damages, plus attorney's fees for willful violations. A disclaimer that the chatbot is an AI does not protect the operator under the bill's current language.
Real estate angle: This is the one to watch if you have an AI assistant on your website or client portal. Real estate is licensed activity in every state. If a chatbot on your site tells a buyer what a contract term means or which financing option to pick, you may be the operator the bill is aimed at, even if you did not build the model. The bill has not passed, but other states may follow New York. Read what your client-facing AI is actually saying, and if it is giving advice, change it.
Source: Holland & Knight
4. Agentic AI emerges as the next phase of proptech
An AI in Real Estate Summit convened in New York City on June 4, drawing 300 real estate and technology leaders during NY Tech Week. The theme was consistent: the industry is moving from AI that generates content to AI that takes action. Agentic systems are being deployed to handle tenant inquiries, predict maintenance needs, schedule repairs, and monitor portfolio performance without a human approving each step. CoStar's Homes AI platform, launched in February, keeps its data within the Homes.com ecosystem and does not use it to train external models.
Real estate angle: An agent that schedules an inspection, sends a notice, or logs a communication is creating records and triggering obligations, which is a different risk category than a chatbot answering questions. Ask every vendor: when your AI takes an action, who is responsible for it? Get the answer in writing before the tool goes anywhere near a transaction.
Source: ICSC
5. Seventy-eight chatbot bills are moving in 27 states
As of this week, 78 chatbot-related bills are alive across 27 state legislatures. New York's legislature passed a five-year moratorium on the sale of AI chatbot toys for young children while an interagency task force studies the risks. The broader wave includes bills on child safety, companion chatbots, professional advice, and disclosure requirements. At the federal level, the Take It Down Act already requires covered platforms to remove deepfake content within 48 hours of a valid takedown request, with compliance procedures due by May 2026.
Real estate angle: A tool that is fine in one state can create liability in another, so compliance is not a one-time checkbox if you work across state lines. Audit what your AI tools say to clients, write down that you did it, and do it again every quarter. If a regulator or a plaintiff's attorney ever comes looking, your paper trail will matter as much as what actually happened.
Source: Transparency Coalition
The theme this week is accountability. Anthropic's IPO filing puts it under public market scrutiny, Microsoft's in-house models give it direct control over its own AI, and state legislatures are deciding, bill by bill, who is on the hook for what a chatbot says. The tools are moving faster than the rules, but the rules are catching up. For real estate professionals, that means two things: know exactly what your AI tools are telling clients, and know who is responsible when they get it wrong. Better to find that out before the first claim than after.
- Jason