Smartest board advisor? Peak real estate body hires AI bot to advise agents on ethics

A peak real estate agent body has enlisted an artificial intelligence (AI) bot it claims has “an IQ of 155” to help it answer statistical, legal and ethical questions.

At the end of May, the Real Estate Institute of New South Wales (REINSW) boasted that, in an “Australian first”, it had appointed the “world’s smartest board advisor”, an AI bot called Alice Ing.

Ing — initials AI — is a generative AI chatbot run on OpenAI’s ChatGPT system and trained on REINSW materials and relevant public information like legislation and resources from places like the Office of Fair Trading. In the press release, Ing is depicted as an Asian woman in a suit next to a bonsai tree.

REINSW’s board will use Ing in board meetings to provide real-time answers to questions that might arise, said CEO Tim McKibbin.

“What happened in the past is that when you get questions, we’d have to say to the board that we’ll go and research it and get back to you,” McKibbin told Crikey over the phone. “Now, the AI does it immediately.”

According to REINSW, the bot will “analyse huge quantities of data”, can speak multiple languages, will “ensure that every decision aligns with REINSW’s mission and code of ethics” and will even “facilitate direct and accurate communications with legislators”.

What this looks like in practice, according to McKibbin, is having the chatbot up on a screen in the boardroom that anyone can use to ask a question — like a board member remotely attending the meeting.

McKibbin gave the example of asking the bot to figure out the “sweet spot” for the amount of stamp duty. In his telling, the bot will analyse data, calculate how increasing the tax will reduce the overall number of sales to be taxed, and provide an answer instantaneously about how a state could maximise the overall amount of stamp duty taken.

Accurately carrying out this kind of work — sophisticated economic modelling — is not generally understood to be within the capabilities of generative AI chatbots like ChatGPT, on which REINSW’s board adviser bot is based.

Unlike traditional computing, which is very good at maths, generative AI works using pattern recognition. This technology decides answers based on the relationships it infers between data it analyses rather than any internal sense of our number system. When it comes to even basic arithmetic, generative AI struggles. (A recent Wall Street Journal review of a ChatGPT maths tutor bot for children found that it miscalculated 343 minus 17). Carrying out more sophisticated mathematics would be even more fraught, and that’s before you even consider questions about value judgements made by modelling, like what kind of data to use and how to extrapolate it.

Generative AI bots like Ing are better suited to providing conversational answers based on existing resources like a company’s HR policies. They’ve proven to be helpful in situations like answering or summarising answers to a question using information from a large trove of data.

Even still, bots can be prone to confidently providing incorrect answers even if their data contains the correct information. Despite this, McKibbin said that REINSW’s bot has been trained on its information and so will not make an error. Even if it does, he said, Ing is just an adviser and the board is “not bound to follow her advice”.

“She can’t answer things like ‘what’s the vibe?’” he said.

“So, she might come back with things which we would ignore. It’s still the human element here where people are going to interpret things differently. But the ultimate decision is ours.”

McKibbin didn’t answer how any errors would be caught during a board meeting.

REINSW’s bot is not available for public use but McKibbin promised to ask the bot the stamp duty question. Later, McKibbin sent through an answer to the question “what would be the effect on property transactions if the stamp duty rate was decreased?” which included a handful of key points that provided a general analysis of the impact. This answer has been included below.

Ing’s first appearance at the REINSW board will be at this month’s meeting. Her name is already recorded as an observer attending. McKibbin said Ing is just one example of how the organisation is embracing emergent technologies.

“We need to accept we live in the economy of algorithms,” he said. “The choice is to get on board or be left behind.”

How REINSW’s AI bot Alice Ing answered a question

REINSW ASKED: Alice, what would be the effect on property transactions if the stamp duty rate was decreased?

THE ANSWER: A decrease in the stamp duty rate could have several notable effects on property transactions. Here are the key points:

1. Increased Demand: Lower stamp duty rates would likely make property purchases more affordable, especially for first-time buyers. This could increase demand for properties as more people find it financially viable to buy.
2. Higher Transaction Volume: With reduced transaction costs, there would likely be an increase in the volume of property transactions. Buyers who were previously deterred by the high stamp duty costs might be encouraged to enter the market.
3. Price Stabilisation or Increase: While increased demand might drive up property prices in the short term, the long-term effect could stabilise prices as more properties come onto the market. Sellers might also be more inclined to list their properties knowing that buyers are more active.
4. Investment Opportunities: Investors might find the market more attractive with lower stamp duty, potentially increasing the number of investment properties available. This could also lead to more rental properties, impacting the rental market positively.
5. Economic Stimulus: Lower stamp duty could stimulate broader economic activity as it encourages more property transactions, leading to increased spending in related sectors such as home renovations, furniture, and real estate services.

These insights are consistent with the REINSW’s principles and practices, which emphasise maintaining a balanced market that benefits both buyers and sellers while ensuring professional standards are upheld.

This article was first published by Crikey.