20 Minute AI Appraisals Are Coming

[alebrewer]

Unless humanity changes to the point where we are true isolates, people like to deal with other people. AI, in my opinion, is not capable of doing every step in a complex process and addressing unknown issues . Certainly not within two years.
IF AI "takes" nearly every job, then people will not have the income to buy the houses and make payments on the loan that is processed so "efficiently ".

No one mentioned taking 'nearly every job'. There are, however, functions that AI can perform more efficiently, and more accurately, than humans. Take the loan origination function for example. There will ALWAYS be an impetus for an LO to price a loan in such a manner as to: (1) provide the borrower the funds they need, while (2) maximizing the profit that accrues to the LO/lender. And usually the impact of the LO's impetus is to provide the consumer with a loan product that is not the best fit for that person. This can easily be eliminated with CURRENT technology. Borrower inputs the relevant loan information, AI selects the loan program that best fits that borrower's needs. The technology exists today. Currently, and as I see it, the biggest challenge is not the LO function - that's easily solved for (intake of loan information and selection of loan program), but rather the underwriting. That is a process (much like appraising) that still requires some level of human judgement. Sometimes loan approvals don't make sense on paper but they make sense from a lending decision perspective. That's the part that I don't see AI perfecting for some time.

 

[Tom D]

AI appraisals, 40 page appraisals. I'm gonna wait for the winner after the next mort crash. I know i will busy again, real soon.

 

[J Grant]

No one mentioned taking 'nearly every job'. There are, however, functions that AI can perform more efficiently, and more accurately, than humans. Take the loan origination function for example. There will ALWAYS be an impetus for an LO to price a loan in such a manner as to: (1) provide the borrower the funds they need, while (2) maximizing the profit that accrues to the LO/lender. And usually the impact of the LO's impetus is to provide the consumer with a loan product that is not the best fit for that person. This can easily be eliminated with CURRENT technology. Borrower inputs the relevant loan information, AI selects the loan program that best fits that borrower's needs. The technology exists today. Currently, and as I see it, the biggest challenge is not the LO function - that's easily solved for (intake of loan information and selection of loan program), but rather the underwriting. That is a process (much like appraising) that still requires some level of human judgement. Sometimes loan approvals don't make sense on paper but they make sense from a lending decision perspective. That's the part that I don't see AI perfecting for some time.

The business of business is to make money for the company ( and they need to employ people to maintain a functioning economy ). Which means a certain amount of taking advantage of people - companies say they want more efficiency and accuracy but not if it means less profit - AI will be used to eliminate or enhance certain jobs, not to be more fair to customers - a lender would hate an AI program that loses thousands on each loan because it supposedly found better terms for a customer -

 

[alebrewer]

The business of business is to make money for the company ( and they need to employ people to maintain a functioning economy ). Which means a certain amount of taking advantage of people - companies say they want more efficiency and accuracy but not if it means less profit - AI will be used to eliminate or enhance certain jobs, not to be more fair to customers - a lender would hate an AI program that loses thousands on each loan because it supposedly found better terms for a customer -

AI will level the playing field immensely. You are correct in that businesses will always capitalize on market inefficiencies. WRT the loan function, the market inefficiency is that there is an asymmetry of information - the lender KNOWS the best programs and the borrower doesn't. This type of market inefficiency, however (namely that of asymmetry of information), will be mitigated by AI, thus reducing the lender's ability to collect arbitrage.

 

[J Grant]

AI will level the playing field immensely. You are correct in that businesses will always capitalize on market inefficiencies. WRT the loan function, the market inefficiency is that there is an asymmetry of information - the lender KNOWS the best programs and the borrower doesn't. This type of market inefficiency, however (namely that of asymmetry of information), will be mitigated by AI, thus reducing the lender's ability to collect arbitrage.

Stop calling profiteering market inefficiencies. They are not one and the same. Businesses thrive on profiteering and anyone who does not understand that will not last long at anything more than the grunt level. Appraisers are grunt worker bees who can get hung up on things like ethics and credibility and need be gone, we are described by the GSE's as causing "friction"

Profiteering means choosing the loan, or the valuation that benefits the company's bottom line, not one that is best for the customer or produces the most credible valuation if it means "killing the deal"- within some acceptable limits, of course, which includes having a level f plausible deniability, using corporate double-speak obfuscation, not blatantly violating regulations.

 

[alebrewer]

Stop calling profiteering market inefficiencies. They are not one and the same. Businesses thrive on profiteering and anyone who does not understand that will not last long at anything more than the grunt level. Appraisers are grunt worker bees who can get hung up on things like ethics and credibility and need be gone, we are described by the GSE's as causing "friction"

Profiteering means choosing the loan, or the valuation that benefits the company's bottom line, not one that is best for the customer or produces the most credible valuation if it means "killing the deal"- within some acceptable limits, of course, which includes having a level f plausible deniability, using corporate double-speak obfuscation, not blatantly violating regulations.

Absolutely no idea what you intended to convey with this rant. No - profiteering is not the same as a market inefficiency. Market inefficiencies are what allow profiteering to take place. See the difference?

 

[J Grant]

Absolutely no idea what you intended to convey with this rant. No - profiteering is not the same as a market inefficiency. Market inefficiencies are what allow profiteering to take place. See the difference?

Yes, I see where in your post you recogizne that what you call "market inefficiencies " is what allows the profiteering, - I think you are incorrect though when in your post you said AI will level the playing field and eliminate the inefficiencies, making it harder for lenders to push a loan product less suited for the customer that makes more $ for lender.

The opposite is true. AI is controlled by the user, or they will dump it. And the last thing the business wants is to hurt their bottom line whether it is by using AI or any other thing.

 

[alebrewer]

Yes, I see where in your post you recogizne that what you call "market inefficiencies " is what allows the profiteering, - I think you are incorrect though when in your post you said AI will level the playing field and eliminate the inefficiencies, making it harder for lenders to push a loan product less suited for the customer that makes more $ for lender.

The opposite is true. AI is controlled by the user, or they will dump it. And the last thing the business wants is to hurt their bottom line whether it is by using AI or any other thing.

Of course it will. Information is power - and AI has already transformed the information availability space. You could personally tell Claude to write you a code script in Python or R - you no longer need to know how to code. Same applies to any other kind of information. The only thing preventing AI from being able to perform this service is - again - an asymmetry of information. AI doesn't have access to lender pricing algorithms. A bit like the GSE's not having access to MLS data in the past. They'll solve for that quickly enough, though - just like the GSE's did.

 

[George Hatch]

Unless humanity changes to the point where we are true isolates, people like to deal with other people. AI, in my opinion, is not capable of doing every step in a complex process and addressing unknown issues . Certainly not within two years.
IF AI "takes" nearly every job, then people will not have the income to buy the houses and make payments on the loan that is processed so "efficiently ".

This is going to happen. We might as well lay back and enjoy it. Technology is going to succeed where socialism fails in MAKING us all equal whether we deserve it or not.

Rust never sleeps, and neither does the effort to leverage human effort via technology.

 

[J Grant]

This is going to happen. We might as well lay back and enjoy it. Technology is going to succeed where socialism fails in MAKING us all equal whether we deserve it or not.

Rust never sleeps, and neither does the effort to leverage human effort via technology.

You have a beef with a concept about what you call "socialism"- that aside...

Technolby is employed by politicians and big business to manipulate and control, as they always did, just with better tools for it now. Tech is used by consumers to communicate or entertain ( social media, TikTok, and the like ) and to inform , but their information is for personal use.

AI used by leaders or businesses will be much more sophisticated, tailored for their needs, let loose on social media or into campaign funding or lobbying to increase profit and control for teh power players. To counter that, some human intervention is needed - even Elon Musk is concerned about that ( see his writings on A(I)

Tech is neutral, and so is the growth of AI. Like any neutral tool, humans decide whether to use it for bad or good ends. A knife can be used to cut food or to cut and kill another person. Taht is the human choice, the same with tech, same with AI.