By Kevin Hawkins with Korey Hawkins | Vol. 1 Post 13
Real AI is a 100% human-created weekly roundup of all things AI in real estate and emerging AI innovations in other sectors likely to impact real estate, posting a new edition every Friday.
GPT-4 Turbo: What real estate pros need to know
Microsoft-backed OpenAI, the creators of ChatGPT, during its first in-person developer event, launched a more powerful AI model dubbed GPT-4 Turbo, currently in Beta, among other advances.
What will these AI improvements do?
- GPT-4 Turbo is said to be a smarter model than ChatGPT-4.
- It will let even non-coders create personalized versions of ChatGPT for individual use.
- Its answers are based on a new knowledge cutoff date of April 2023.
- It can handle up to about 300 pages of content – or 100,000 words – in a single prompt, vastly expanding the capabilities of current versions of ChatGPT.
- For developers, the cost is 2-3x cheaper.
- GPT-4 with Vision supports DALLE-3 and Text-to-Speech (TTS), which are all now in the API, meaning it can accept an image as part of a prompt and provide an appropriate textual response, generate images, and it can be spoken to and then respond using its voice. Oh, and there are six different voices you can choose to hear its answers.
- OpenAI also announced its significantly improved speech recognition model called Whisper V3, which can handle dozens of languages.
We are already seeing the impact of computer vision on real estate as nearly the entire MLS platform is integrating with industry leader Restb.ai.
If you haven’t seen the Restb.ai video showing how an ADA and Fair Housing-compliant entire property listing can be completely generated immediately after the agent uploads the property photos into their MLS, click here. You will see in real-time a concrete example of how computer vision will make life as an agent easier and more productive.
But this is only the beginning of extracting data and intelligence from images. The new GPT-4 Turbo with Vision demonstrations shows how incredible its AI is in instantly identifying individual items inside a photo and explaining what they are. Check out this video on X from Robert Lukoshko to see what we mean.
Creating your own GPT is a breakthrough previously reserved for enterprise clients and developers. Open AI now claims, “Creating one is as easy as starting a conversation, giving it instructions and extra knowledge, and picking what it can do, like searching the web, making images or analyzing data.”
In terms of AI-powered multi-lingual tools, these might empower more potential non-English speaking home buyers to become more educated and knowledgeable about the real estate market and the benefits of homeownership. These new tools could also enhance agents’ communications and outreach capabilities.
The biggest takeaway is the OpenAI announcement showing the direction AI is going (and we’ve been forecasting): better TTS. Having to type in prompts in an AI world makes absolutely no sense. We can talk to Siri, Google Nest, and Alexa, but we need to speak to ChatGPT.
Finally, CNBC noted that OpenAI offered its own guardrails for AI, saying it will step in and defend customers and “pay the costs incurred if you face legal claims around copyright infringement,” echoing what Google, Microsoft, and Adobe have said.
AI Five Fast Facts
Survey of 216 college professors from 67 of the top 100 US computer science programs:
- 56% of computer science professors at top US research universities surveyed described the corporate leaders as “extremely disingenuous” or “somewhat disingenuous” in their calls for regulation of AI.
- One in four believe AI will become so advanced at medical diagnoses that it will generally outperform doctors.
- 85% of survey respondents said AI can be at least somewhat effective in predicting criminal behavior, but only 9% said it can be highly effective.
- 62% said that misinformation is the biggest challenge in maintaining the credibility and authenticity of news in an environment that includes AI-generated articles.
- 95% of those assessed described AI’s current deepfake capabilities as “advanced” when it comes to video and audio content.
Source: Axios, Nov. 8, 2023, with Generation Labs and Syracuse University.
AI Headlines Take 5
1. YouTube creators will soon have to disclose the use of gen AI in videos or risk suspension | Associated Press 11/14/23
YouTube’s updated policy will allow the removal of AI-generated video that simulates a person’s voice.
2. 5 Steps Your Business Needs to Take to Build a Responsible AI Program | Inc. 11/14/23
Maximizing the use of AI is essential, but so is taking precautions for any AI pitfalls.
3. How the Real Estate Industry Can Effectively Leverage AI | Urban Land Magazine 11/13/23
Maurice Conti pines on the impact of Generative AI in real estate development. (Requires ULI Guest Account signup.)
4. LinkedIn Introduces New AI-powered Premium Experience | Inman 11/1/23
Are you looking for a reason to pay LinkedIn? Premium users now get access to an AI job coach to assist with job hunting, including answering questions about a company’s hiring process and helping determine if they’d be a good fit.
5. How to Use ChatGPT to Plan Your Winter Travel | Kimpton Hotels 10/8/23
Yes, this is a self-interested post, but solid ideas on how to use ChatGPT as a travel planner assistant.
Quote of the week