AI: Ensuring the Integrity of Elections & Voting

by | Dec 7, 2020 | News, Thought Leadership

Ensuring the Integrity of Elections & Voting:
Artificial Intelligence, Data and Advanced Technologies

Why are advanced technologies such as Artificial Intelligence & Blockchain not used to ensure Voter Integrity and Security?
Severence M. MacLaughlin, Ph.D
Managing Partner
DeLorean Artificial Intelligence

*Disclaimer: I do not purport to be a legal or election expert

I, like most Americans, am very much exhausted by the constant digital barrage of 2020 Election. We had hoped it would end on evening of November 3, 2020 with a smooth continuation or transition based on the legitimate winner of the Presidential election, but alas it continues. The integrity of the 2020 Election and voting in general has been a constant topic throughout the past twelve months, mostly focused on the security surrounding Mail-in ballots. Now, there are serious challenges associated with voting integrity and voter fraud. Are these allegations true – I honestly do not know. What I do know is that the regular American probably shares my opinion that we would like to have faith that elections in the United States of America are fair and secure, and further that every legitimate vote and voice is counted.

There are allegations of widespread voter fraud and breaches of election security, which come in a number of variations:
1. Votes cast by people who are dead
2. Unusually high voter turn out
3. Voter turnout higher than voter registration or a population of a geographic area
4. Missed votes
5. Fraudulent use of Mail-in ballots
6. Not proper signature checks
7. Not proper address checks
8. People voting in multiple locations
9. Ineligible individuals voting
10. Backdating of Mail-in ballots

It does not matter where on the political spectrum one might fall; all Americans would logically desire a system that alleviates the aforementioned opportunities to malign the integrity of the election process. Additionally, this is not just an American issue as the digital transformation progresses globally and into the elections of other countries the issue of election integrity and security is a challenge faced by all democracies.

What I do not understand is why we as a society have the technological capabilities to ensure that these ten variables (which are not exhaustive) are moot through a thoughtful scientific design of a national voting integrity system. I do realize that each state and county are responsible for their respective elections, but our Federal and State Governments do have a fiduciary responsibility to ensure this desired integrity.

Further, if Apple can ensure the security of every single one of their customers’ digital privacy and security, why can our government not ensure that for our elections? Security at a level of Apple as deployed through their Apple iCloud accounts would engender confidence and faith in a fair election process. I am using Apple as an example, but Apple has better digital security than my bank; I would argue that people are probably more afraid of financial fraud than a digital breach of their iPhone. Why are we as a society so slow to adopt the technologies developed by our scientists to ensure the continuity of not only our electoral process but those issues or daily concerns such as banking and healthcare. I do recognize that the COVID-19 pandemic has been an evolutionary event in digital transformation of the globe, however, as I watch our mainstream media, I am astonished that we as a nation have not employed technologies that could easily solve these issues or concerns – now implementation may be more difficult but that is just heavy lifting.

Why have our respective government agencies responsible for the security of elections not adopted tools such as artificial intelligence and block chain to alleviate some of the inherent opportunities for fraud in the electoral process. How could someone in this day and age use a deceased person’s identity to register to vote or vote?? Why are there not scientific processes in place to ‘red flag’ counties or precincts that experience out of norm voter turnout or voting?

My main conclusion as I think about this topic riding a bike, walking the dog or just prior to going to bed is that we have the technology, however, we do not have the big data infrastructure to harness all of the required sources of data that would need to be fed into an Artificial Intelligence solution.

My suggestion is that either the Federal Government (Federal Elections Commission, Dept of Homeland Security) or collaboration between State and Federal electoral agencies create a System of Intelligence (SOI, AI) that has the ability to harvest data (structured and unstructured data that can ensure individuals who are eligible are able to vote and vote only once. But bringing those data sources together would be the most challenging for a couple of reasons: 1. Privacy issues – but the government already has this, 2. Inter-agency data sharing – yes you would assume our government would have one true source of data, but that is another diatribe to write.

What would the solution look like and data sources (by no means perfect or exhaustive, but a whole lot better than was is current).
1. Eligible Donors
a. Data Sources:
i. Voter Registration data base
ii. Social Security list – has both live and dead names and addresses.
1. People want their money so addresses would most likely be correct
2. Having just lost a family member recently I do know there is a pretty good process of logging the death of an individual
iii. Medicare list – have records of live and dead individuals
1. Healthcare is important to people so their addresses would be up to date.
2. If someone died there is a record
iv. Postal Address list
1. As we have seen in the 2020 election this could be fallible in some instances but could be cross checked with Social Security and Medicare.
v. Social Support Benefits list
1. Again, people make sure that they receive their benefits so addresses would most likely be up to date.
vi. Internal Revenue Service lists – the IRS is good at their job and I would bet their demographic information for tax payers is pretty decent.
vii. REAL Compliant State Issued IDs – you cannot fly without these…
A Real ID-compliant form of identification requires the following pieces of data:
1. Full legal name
2. Signature
3. Date of birth
4. Gender
5. Unique identifying number
6. Principal residence address
7. Front-facing photograph of the applicant

Combining these seven data sources into one true source of data and environment that could be referenced at the point of voting or request for a ballot would allow ONLY eligible donors in their respective geographies admittance to vote or access to vote.

Besides the data that would be required to reference to ensure someone was eligible, the other large problem is linking the voter to the ballot. Bar codes, QR codes and REAL compliant identification are the answer. Again, this is not rocket science.

A. In person voting: When a person attends a polling location their identity is confirmed by scanning their REAL compliant identification card. This identification card downloads the aforementioned data and cross checks with the System of Intelligence using RPA and other data science tools to ensure that they are registered to vote, they are eligible in that location and that they are alive. Then, the REAL ID is linked by scanning the ballot bar code or QR code. Now the voter is validated as well as digitally linked to their ballot. Further, using computer vision a photograph and or signature can be validated.
B. Mail-in and Absentee Voting: When a person requests either an absentee ballot or Mail-in ballot they must provide the REAL ID unique identifier number. This number is then digitally linked to the ballot that is sent to them. When the ballot is returned, the voter is confirmed. This would allow for only eligible voters to be given these types of remote ballots as well as prevent double voting of remote and in person voting. Further, if there were discrepancies the digital link could be confirmed for recounting purposes.
C. Increased voter turnout – this system of intelligence would allow red flagging of any statistically differential event in terms of number of voters that would be identified as potential security breach and would require human intervention. The key here is that every ballot, vote and voter is digitally linked and pre- and post-checked for eligibility.

The utilization of AI with such a massive data base would not only support the security and integrity of the election but also allow prediction where issues may be, as well as used to plan for future elections in terms of locations, number of voting booths and personnel.

Let’s take this one step further, blockchain. Blockchain has been heralded as a solution for centralizing and accessing personal health information in a very secure manner. If blockchain technologies and or vendors can provide a secure method for the portability of healthcare information (which is a huge achievement and great for increasing health outcomes so that all HCPs have access to your medical history), why can’t we do the same for you, a citizen’s voter eligibility and credentials. Just show up at the polling location scan a QR code and you are immediately identified and can vote? Maybe Starbucks could help….with the scanning QR platform.

Now I know what you are thinking, this will cost a fortune – well yes but not more than what has been spent by local and federal governments on this election. Georgia spent $100M on voter equipment in 2020 and 5 other states had similar budgets. In 2002 in response to the 2000 election the Federal Government spent $3Billion, so we would not have the “hanging chad” issue. I can guarantee you that a technology company partnering with a Systems Integration company could achieve a System of Intelligence as outlined above for $100M plus run and maintain costs. For $500M we would have one rock star of a security system.

Think about this when we discuss that $14Billion was spent on the 2020 Election.

In conclusion, do we have the technology to achieve election integrity and security – YES. Can we afford to do it, YES. Will it be easy, unfortunately NO; but nothing worth doing is easy.