I just listened to a podcaster I’d stopped listening to some time ago. Dr. Mark Calabria, former Head of FHFA was interviewed. Yes, he is selling a book (which I’ll now likely buy after hearing him speak). I approached it with a certain amount of skepticism.

He has as thorough an understanding of the issues we face, as a non-appraiser can have. Well worth the 48 minutes. Then decide for yourselves..

Apparently, the message focus has been lost.

Why not? You’ve done the research, narrowed down the search results, selected appropriate comps. That’s your market, a micro view of it at least. If your adjusted values are in alignment after such reasonable adjustment applications, then you’ve achieved market reaction level adjustments. Market value is a moving target.

A logical fault to believe in the ability for data to be meaningful when extracted through regression tools and such, but then to deny the reality that all those data points come from individual homes, and one could simply just review individual homes for meaningful data. If one trusts the regression to be so accurate for market extraction for feature differences like a bed is worth this much, parking that, then one has no room left for quality of materials considerations. A balance between the two is wise. Problemo?

Special appraisers life hack; One can extract the market reaction to be used in the adjustment grid for pretty much anything home related, by knowing how much a contractor would charge for that completed service including materials. Then estimating down for depreciated effective age.

Every single value adjustment one could need to extract for individual adjustments, can come from just the simple method of closely examining a small set of reliable comparables themselves while filling out the grid. No special special proprietary software or advanced mathematical skill is necessary. The market reaction is already present, in the market data, for those real world homes that real world people already showed market acceptance or market resistance regarding.
The closer the comparables selected are to each others character and effective age, there is a diminished need for the amount to be adjusted. It’s the ‘distributed allotment’ method. A regular joe keeping up in uniform suburbia knows how much it costs for all those individual components, they have to actually pay for them. Most minor item adjustment amounts can be rooted in real world costs for better report defensibility than pointing to some chart and saying the extracted market data told me so. I got a sweet deal on a hot water heater because I paled around with the owner, and tipped him a twelve pack of beer last time around. Positive value gain, achieved. Winning. And saving. Two distinctly different consumer expectations, gaining value vs saving money.

The garbage in garbage out argument and data becoming stale argument, that’s happening right now in real time. There is no such thing as artificial intelligence, although we do seem to have software which seems to be capable of self learning, while being fixed in time to comply with it’s initial operating peramiters, however flawed they may have been. Reminds me of the ending of Tron; I did everything, everything you asked. I know you did. I executed the plan. As you saw it. You promised that we would change the world together, you broke your promise. I know, I understand that now. I took the system to it’s maximum potential. I created the perfect system. The thing about perfection is it’s unknowable, it’s impossible but it’s also right in front of us the whole time. You would not know that because I didn’t when I created you. I’m sorry.

If the AI system was so good, why did FNMA suddenly limit the scope of it’s use, yet continue to sell this product as an adequate replacement for human appraisers?

https://appraisalbuzz.com/real-prompts-for-chatgpt-for-appraisal-reports/
https://appraisaltoday.com/2023/05/19/chat-gpt-for-fannie-form-appraisal-reports/
https://www.hypewired.com/real-estate-ai-appraisals/
This third one is a peach.

These blind idiots, they’re all in on this new tech. Just one problem, the AI system relies on the AVM data. Optimism vs pessimism is an irrelevant concept when talking about this technology. The potential for systemically increasing data irregularities and faults to one day dominate and skew data perception beyond any hope of salvageability or reliability is astonishing. So is the potential for malicious actors to affect these data outputs and data input systems reliability. It’s like these fools have never even bothered to read security tech news, which continues to disturb with the breath taking pace of malicious actors advancements. Smart technology is so very very stupid and poorly developed with consistently identified vulnerabilities no matter where you look. Yet most people continue to believe the public relations lines of for profit companies regarding security and competent tech development. They continue to pay hard earned dollars to actually bring more of this technology into their lives rather than working diligently to mitigate the risks and return to a safer more limited technological usage footprint. Which would be the wiser approach these few decades later now that we can so clearly see how exploitative and exploited all these technical advances have become.

They just hand it off to the IT department with blind faith and hope for the best. Meanwhile routine penetration attempts worldwide continue. Search this term; Hacks per minute worldwide. It’s the human error factor, accentuated with government bureaucrats.
https://www.zippia.com/advice/cybersecurity-statistics/

One ponders how the appraisal robots of the future will perceive health and safety hazards, or if perhaps we’ll see a funny video of them slipping on a dog turd in the living room of an abandoned double wide, totally short circuiting and smoking out. Dare to dream. When the FNMA CU system eventually goes into the wrong hands, it’s going to be big. I mean, similar events have happened before. The miracles of working with third party tech companies. Especially all these start ups and amc’s working ‘appraisal modernization’. The exposure by simple company count is tremendous, there are now thousands of companies networking and relying on each others tech and each others security, to accomplish what used to just take a manual effort with a few specialty licensed vendors, a lender, and a few people within the gse networks. Now they’re addicted, they bring new third party companies into these networks on a near daily basis. Bureaucrats often find it difficult to stop the process of handing out special favors, when they do so with other peoples money, no risk to themselves personally.

https://www.theguardian.com/business/2017/oct/10/deloitte-hack-hit-server-containing-emails-from-across-us-government
The Guardian has established that a host of clients had material that was made vulnerable by the hack, including: “Fannie Mae” and “Freddie Mac”, the housing giants that fund and guarantee mortgages in the US. “The hackers had free rein in the network for a long time and nobody knows the amount of the data taken,” said one source. * “A large amount of data was extracted, not the small amount reported. The hacker accessed the entire email database.”

Hell, most of the people here have already had your information pilfered from a company with lackluster security, but may simply be unaware it happened. Check your credentials; Have I been Pwned. If you get the green screen instead of the red screen, consider it a small miracle.
https://haveibeenpwned.com/

The important moral lesson here; You can’t substitute some technical program or software application for the reliability and security which comes from a hard days work, completed by a qualified and competent human being. These psychopaths dreams of some automated utopia will always remain just somehow so slightly beyond their grasp. Something they can envision and imagine, work towards their entire lives, constantly make advancements towards, but never actually successfully implement as planned. AI will result in more harm than good, the first roll out stages are already a disaster, as people find unique, novel, and quite unexpectedly surprising ways to exploit the technologies capabilities. Automating these systems when the entire world seeks to exploit Americans technical systems vulnerabilities, and they’re getting quite skilled and proficient, successful in their goals, such is a fools errand.

https://www.fhfaoig.gov/sites/default/files/WPR-2022-002.pdf
Related FHFA report on use of AI/ML technology.

In June 2021, Fannie Mae described its use of AI/ML at the time as limited,”
and it leveraged AI/ML in applications such as property valuation, fraud detection, post-purchase loan reviews, and language processing. In May 2022, Fannie Mae reported to OIG that the Enterprise also uses chatbots to answer customer questions regarding the mortgage process and has since retired AI/ML use in post-purchase loan reviews
&
Use of AI/ML through third-party providers: The Enterprises rely on third-party cloud providers to meet AI/ML’s data storage and computing needs. Additionally, the Enterprises use third-party machine learning platforms for data processing and model development. These third-party uses present additional risk concerns, including information security, business resiliency, and concentration.1

Well they were using artificial technology for review purposes, if I’m interpreting that accurately. Is what’s coming through now the delayed result? Hell that does not seem to bother them when the amc industry is using this tech, sweetheart deals and all which result in continued abuses of the system where amc’s discriminate against appraisers and exclude us from participation in our own industry, while ramping up the use of this very same automation. aka; Restraint of Trade Practices.

If there is more in this doc, not sure, I just skimmed this one, a random find. I googled this term; Fannie Mae Artificial Intelligence. There is a lot there…

https://www.findlaw.com/smallbusiness/business-laws-and-regulations/restraint-of-trade.html

Oh no, getting closer to home, that’s the first I’ve heard of CO.

Actually, have what may be a relevant article on the matter, with legal defense implications, if an AI based program was used to make these decisions. Some college professor tried to use an AI system to recognize if students were cheating paper writing assignments with AI, which is becoming a real problem lately. It came back that all were likely cheats, which was not true. One student then ran the professors older works through the same program and the AI said yep, that’s a cheat. When technically illiterate people place faith in biased programmed software that is not quite as competently developed as they think. Your suspicions are well warranted and entirely plausible. I mean, they’ve only so much as stated they want this technology to become utilized more broadly for that exact purpose.

https://www.fanniemae.com/research-and-insights/perspectives/how-will-artificial-intelligence-shape-mortgage-lending

Someone should do an experiment and see if these AI programs can write up an appraisal properly or not. After all, the GSE’s are touting the AI integration as ground breaking for their review and accompaniment purposes in dealing with appraisals, which seems flimsy if it could not properly perform the activity of fulfilling the requirement of making over 2,000 important quality decisions for each and every appraisal report developed. (Remember that article where the appraiser said we make that many important decisions every time we fill out the 1004?) And you know what, that gets me thinking. Imagine this scenario; a PAREA graduate getting the lions share of amc requests, while using an AI program to write up the ‘appraisals’. I am in agreement we need to completely shut down all AI based programs and shelve them permanently.

https://www.rollingstone.com/culture/culture-features/texas-am-chatgpt-ai-professor-flunks-students-false-claims-1234736601/

Also the incandescent light bulb sales ban comes into effect in July. This is the very last month to get them if desired. I just paid another five hundred dollars for incandescent. Found the only seller on ebay whom had legit 100w reveals, at 170 per 48 pk. $3.50 per bulb but one day people may look back and say, wished I would have had more. These other types disrupt sleep patterns, I switched all back to incandescent and am sleeping much better lately. These things are appreciating faster than gold. Only the government could screw something up this badly. (That’s my new favorite tag line, borrowed from another appraiser here, brilliant.)
https://www.ebay.com/str/wholesalelightbulbliquidators?_pgn=1

In reading over arguments for another bogus complaint they sent out (to Colorado) Im beginning to think they are using AI chat to communicate their fraudulent complaints; or should I say bogus complaint in support of their banking fraud exercises?

Now every time I read an AI based article, my mind goes to the Fannie Mae Fraudulator. Inspired title. Mr Ford, you’re a very busy person, we sincerely appreciate your time and efforts.

The state sent me the complaint letter from Fannie. The only statement from Fannie was “Use of dissimilar comparable sales.” I sent in my work file as directed back in August 2022. Believe me, I haven’t been losing any sleep over this.

Pat (#4) that means a lot coming from you!

(For readers who don’t already know, Pat T. has the oldest surviving appraiser’s license in his state. When he first started as an appraiser the worst thing appraisers risked was being attacked by any saber-toothed lions that might be in the cave being appraised; or hitting their thumb with the rock chisels used for writing reports on the walls. )

In 2018 Pat was awarded a Lifetime Achievement Award in San Antonio, TX by his peers from all over America. I’m truly humbled by, and similarly proud to be his friend. Much respect Amigo! (though I will carefully avoid showing it in public, lest people begin to talk).

Claiming exemption from USPAP is not actually the same as BEING exempt. Since FNMA doesn’t conform to their own published standards (or any other recognized standards), I contend they do not meet the requirements under FIRREA for exemption.

Facilitating fraud against a federally insured bank is not an accepted standard anywhere, outside of FNMA.

I read the HousingWire article.

I’m also told very few have any experience outside of Texas, and none very recent. Several sources they rely on lack credibility for the purposes they are being used. Neighborhood Scout left nearly 50% of my own neighborhood out when they mapped it. (I was auditing its reliability).

It’s pretty clear that screening is going on to make sure only low-hanging fruit is selected to fulfill buyback quotas. Small, midsize banks can least absorb the hits.

Thank you, Brian.

If they filed with the state board you should have heard the details of the complaint already. Are you referring to a complaint they made to the bank where the bank is supposed to reach out to you for a rebuttal?

OR was it a state complaint that already hit the board and you responded and are waiting for the state’s decision on the complaint? If it’s the latter, all you can do is wait as patiently as you are able. If yu think it may help you are more than welcome to send a copy of this article as an indicator that many complaints (50%) are quota driven.

Former FNMA Subject Matter Expert