'Alice Ing' charges only $25 per month to perform its role as a board advisor for an Australian real estate institute, despite having a photographic memory, an “encyclopaedic knowledge” and speaking multiple languages.

Alice is said to be the first artificial intelligence bot (hence its initials, AI), to be "appointed" a board advisor for an Australian organisation.

The bot also comes with its own AI-generated image of a woman of South-East Asian descent wearing a business suit and posing near a bonsai plant.

The bot, which runs off OpenAI’s popular ChatGPT system, is an advisor to the board of the Real Estate Institute of New South Wales (REINSW), a peak body for real estate industry workers.

The appointment, announced at the end of May, follows a small number of international companies providing AI assistants to their boards and leadership teams.

Abu Dhabi-based conglomerate International Holding Company also appointed an AI board observer in May, naming it Aiden Insight (again with the initials AI) and asking it to “enhance decision-making, risk management, and strategic planning”.

REINSW says its Alice bot will help it more quickly provide information to its board members and assist in its own decision-making processes.

CEO Tim McKibbin tells Information Age that businesses “need to see the positives, and guard against the negatives” of AI — and the drawbacks he alludes to can be numerous.

Can you trust an AI advisor?

Alice’s creators say their system has been fed data such as state legislation, regulatory information and details of REINSW meetings, statements and intellectual property.

They say Alice can also search an intentionally limited number of websites for additional information, because they are relevant to its role.

Alice's web access has been limited because even the latest versions of Large Language Models (LLMs) like ChatGPT and Google’s Gemini, which have wider access to the web, can produce what are known as hallucinations — when AI systems generate incorrect, incoherent and sometimes dangerous information.

Scraping the ocean of content on the open internet has seen LLMs take jokes on Reddit and satirical news headlines as fact, while some have even been tricked into ignoring their own rules or made to provide illegal advice.

“If there is conflict between what Alice finds on the web and what she has in her memory bank, then the memory bank stuff trumps it,” McKibbin says.

He maintains that the advice Alice provides will always be checked by the institute, just as a human advisor’s would be.

“The board will consider that information — it doesn't mean they're going to slavishly follow Alice’s recommendations,” he says.

“They will bring other factors to the table, and they will consider what else is provided, along with other matters.”

In an AI-generated statement, Alice says it will ensure every board decision “aligns with REINSW’s mission and code of ethics, calling out any potential inconsistencies or conflicts”.


The Real Estate Institute of NSW says its Alice bot will provide its board with crucial data and information. Image: Shutterstock

‘A PR stunt’?

Toby Walsh, a member of the federal government’s AI Expert Group and the Australian Computer Society's AI Ethics Committee, tells Information Age that he sees the introduction of Alice Ing as “a terrible idea” and “a PR stunt”.

While he says it is “wonderful to see boards being more evidence-based and having more technical expertise”, he has “serious concerns” about ChatGPT’s track record of making things up.

In its latest report on the state of AI, management consulting group McKinsey & Company found nearly one-quarter of 1,300 survey respondents said their organisation had experienced negative consequences from generative AI’s inaccuracy.

Almost half of respondents (44 per cent) said their organisation had experienced at least one negative consequence of generative AI, such as issues with data privacy, bias, intellectual property infringement, security or incorrect use.

Inaccuracy was found to be the most common impact on organisations, followed by cyber security and explainability issues.


Source: McKinsey Global Survey on AI, February 2022 - March 2024

Walsh says an AI bot is also unable to hold people to account like a human can, and it cannot watch out for inevitable but unpredictable occurrences known as ‘Black Swan’ events, including cyber security issues.

“AI is terrible at watching out for Black Swan events, because they can only do what is in the training data sets. If it’s a Black Swan event, by definition it wasn’t in the training data,” he says.

Despite these risks, the McKinsey & Company report also found businesses were starting to see benefits from the use of analytical AI systems, including cost savings and revenue increases from AI use in marketing and sales.

‘Why wouldn’t you do it?’

REINSW’s CEO said he was confident Alice would perform well and increase the efficiency of the organisation’s board.

“The question for me is not so much, ‘Why would you do it?’ I think the better question is, ‘Why wouldn't you do it?’” Tim McKibbin said.

“Because Alice has an IQ of 155, which makes her the smartest person in the world.

“Alice can take data sets that we give her and she can cross-reference and crunch up all the numbers and deliver information to us, literally at the speed of light.

“Why wouldn't you want somebody like that on your board?”

Walsh admits the technology behind Alice does have some positive uses.

“If the board wants to brainstorm about different business strategies that they can explore, then actually tools like ChatGPT are really good — to help you not to do the actual brainstorm, but to see the ideas which then humans can take on,” he says.

McKibbin said he originally hoped to put Alice on REINSW's board in a non-voting capacity, but the organisation decided not to after seeking legal advice and holding internal discussions.

“I can see that happening in the not-too-distant future,” he said.

McKibbin added that concerns about AI bots taking over jobs previously done by humans was “a natural concern in times of enormous shakeups”, and pointed to previous technologies freeing people up to take on less menial jobs.

“I think the same thing will happen here. But do I think there is going to be an upheaval? Look, absolutely. There is going to be a shake up here, I think undoubtedly,” he said.

“But you're not going to be able to hold it back, is the short answer — it's going to happen.”

Julian Moore is AI chief for Strategic Membership Solutions, and helped REINSW implement Alice.

He says Alice’s appointment “signals a new era in which AI not only supports but actively participates in Australian leadership and governance”.

"I think rather than viewing it as a board member, view it as a tool that board members can ask a question of to make a faster decision based on the data applied," he says.

“Even the Australian Institute of Company Directors say that all board members should be fully aware of all information before making a decision — Alice allows that to happen fast.”

Moore says he has been contacted by many other organisations who are looking at implementing similar technologies.

“It's something I think thousands of more people will do,” he says.

“And once they've seen how straightforward it is, they'll make their own.”