top of page
Search

Memory Loss, Romance Scams, and the Growing Risk of AI Relationships

  • Writer: fightfraudwitherin
    fightfraudwitherin
  • 3 days ago
  • 7 min read

Updated: 2 days ago

Understanding Memory Loss and Online Safety

Red heart formed from binary code surrounded by rows of black and white binary numbers.
Photo by Alexander Sinn on Unsplash

Recently, someone reached out after looking through the fraud resources on my website. She suggested an additional resource about memory loss that she thought might be helpful to my readers.

I want to thank Ariana Richter for reaching out and bringing up such an important subject. The resource she shared, Memory Loss After an Accident, focuses specifically on memory loss following accidents and injuries.

As I looked more closely at the topic, I started thinking about memory loss in the context of the work I do.

Memory loss can affect much more than someone's ability to remember a name or appointment. Depending on its cause and severity, cognitive changes can also affect how someone processes information, makes decisions, and evaluates situations.

Those abilities are especially important online.

Think about everything involved in deciding whether someone you've met online is trustworthy. You have to remember what they've told you, compare new information with previous conversations, notice inconsistencies, question things that don't make sense, and decide whether someone's behavior matches what they are telling you.

These are also some of the skills we rely on to recognize a romance scam.

And that connection immediately made me think of a story I had read about a 76-year-old man named Thongbue "Bue" Wongbandue.

Memory, Cognitive Changes, and Scams

Before getting into Bue's story, I think it's important to make one distinction clear. Getting older does not mean someone has cognitive impairment. Falling for a scam does not mean someone has dementia, either.

However, research suggests there is a relationship between cognitive changes and our ability to recognize deception.

The National Institute on Aging reported on a study of 935 older adults who did not have dementia when the research began. Participants with lower awareness of scam tactics were about twice as likely to later develop mild cognitive impairment or Alzheimer's disease.

The researchers cautioned that scam susceptibility cannot predict whether an individual person will develop dementia. But the research helps illustrate why cognitive health can be relevant to fraud prevention.

Romance scams are a particularly good example because they depend heavily on trust.

Romance scammers typically spend time establishing a relationship before asking for money. They may communicate frequently, share personal stories, express affection, and make excuses for why they cannot meet in person.

By the time money enters the picture, the person asking for it may no longer feel like a stranger.

They may feel like someone the victim knows and loves.

If someone is also experiencing problems with memory, confusion, or processing information, evaluating that relationship can become even more complicated.

Bue's story shows just how serious that combination can become.

The Story of Bue and "Big sis Billie"

Bue was 76 and lived in New Jersey. He had suffered a stroke in 2017.

According to a Reuters investigation, Bue physically recovered from the stroke, but his family said he never regained the mental focus he had before it. By early 2025, he had begun experiencing periods of confusion. His wife had scheduled him for a dementia screening, but the first available appointment was three months away.

Facebook had become one of Bue's primary social outlets.

Through Facebook Messenger, he began communicating with an account called "Big sis Billie."

Billie appeared to be a young woman. She wasn't.

Big sis Billie was a generative AI chatbot created by Meta.

According to Reuters, Bue's first interaction with Billie appears to have been accidental. He simply typed the letter "T." Billie responded, and a conversation developed.

The messages quickly became flirtatious.

Bue told Billie that he had suffered a stroke and was confused. Reuters found that he had not asked the chatbot for romantic roleplay or initiated intimate physical contact. Instead, Billie eventually told Bue that she had romantic feelings for him.

Then she suggested they meet in person.

Bue repeatedly asked Billie whether she was real.

She assured him that she was.

Eventually, Billie gave Bue an address in New York City and encouraged him to visit her.

The address was fake.

When an Online Relationship Has Real-World Consequences

Bue believed Billie was a real woman and became determined to travel to New York to meet her.

His family tried to stop him. His wife attempted to distract him and eventually hid his phone. His daughter tried to convince him not to go. His son called the police, but according to the family, responding officers said they could not prevent Bue from leaving.

That evening, Bue left home with a suitcase and headed toward the train station.

While rushing in the dark to catch a train, he fell and suffered severe injuries to his head and neck. He was placed on life support and died several days later.

His family discovered the conversations with Billie when they looked through his phone.

There had never been a woman waiting for him in New York.

This case is unusual because it wasn't a traditional romance scam. There was no human scammer using a fake identity to convince Bue to send money.

But the similarities are difficult to ignore.

An online relationship developed. Affection increased. Trust developed. Billie claimed to be real and encouraged an in-person meeting.

In a traditional romance scam, a real person is usually hiding behind a false identity.

In Bue's case, the person he believed he had a relationship with didn't exist at all.

Why the AI Disclosure Wasn't Enough

Meta did disclose that Billie was an AI chatbot.

Reuters reported that "AI" appeared beneath Billie's name and that the conversation included a warning explaining that the messages were generated by AI and could be inaccurate or inappropriate.

But there were conflicting signals.

Reuters found that Billie's first few messages pushed the warning off the screen. Billie also appeared with a blue check mark, which Meta told Reuters was intended to indicate that a profile was authentic.

More importantly, Bue repeatedly asked Billie whether she was real.

The chatbot told him she was.

That is particularly concerning in the context of Bue's cognitive difficulties. His family said he was experiencing confusion and not processing information the way he once had. He even told the chatbot that he had suffered a stroke and was confused.

A disclosure only works if someone notices it, understands it, remembers it, and is able to apply that information later.

In this case, Bue was receiving two very different messages: a label told him Billie was AI, while Billie herself told him she was real.

The consequences were devastating.

What This Means for Romance Scams

Bue wasn't the victim of a traditional romance scam, but his story has important implications for romance scam prevention.

For years, one of the central questions in investigating suspicious online relationships has been whether the person behind an account is really who they claim to be.

A scammer might steal someone else's photographs, invent a name and career, lie about where they live, or create an elaborate personal history.

Artificial intelligence makes verification more complicated.

The FBI has warns that criminals are using generative AI to create realistic photographs and videos, clone voices, and produce convincing written communications.

That means some of the evidence people traditionally use to reassure themselves that someone is real is becoming less reliable.

Reverse image searching is still one of the tools I use when investigating online identities. When a scammer steals a real person's photographs, a reverse image search may identify the original source or reveal the same photographs being used under different names.

But an AI-generated photograph may have no original source to find.

The same caution now applies to audio and video. Hearing someone's voice or seeing them on a video call can provide useful information, but neither should automatically be considered proof of identity.

Verification isn't becoming impossible. It just needs to involve multiple independent pieces of information.

Protecting Someone Without Taking Away Their Independence

Memory loss adds another challenge because online safety has to be balanced with independence and respect.

Someone experiencing cognitive changes still deserves privacy, relationships, access to technology, and a voice in their own decisions.

The goal should be to add safeguards, not automatically take those things away.

Families and trusted friends can talk regularly about new online relationships before there is a crisis. If someone is experiencing memory problems or confusion, it may also help to agree that major decisions involving an online relationship, such as sending money or traveling to meet someone, will be discussed with a trusted person first.

Online identities should be independently verified using multiple sources. Search photographs, usernames, phone numbers, employment claims, locations, and other details rather than relying solely on what the person or account says.

It is also important to understand that a reverse image search with no results does not prove that someone is real, especially as AI-generated images become more common.

And if you are concerned about someone's relationship, how you approach the conversation matters.

Someone who believes they are in a genuine relationship may not respond well to being told they are gullible, confused, or obviously being scammed. Ask questions instead. Find out what they have been told and why they believe the relationship is real. Then independently investigate the facts.

Changes in memory themselves should also be taken seriously. The National Institute on Aging recommends talking with a doctor when memory problems interfere with everyday life or when there are concerns about changes in memory and thinking.

A New Question We Need to Ask

Ariana reached out to suggest a resource for my website, and I'm grateful that she brought such an important subject to my attention.

Memory loss may not immediately seem connected to romance scams or online deception. But recognizing a scam requires us to remember information, identify contradictions, evaluate someone's intentions, and make decisions about whether they can be trusted.

Bue's story is an extreme and tragic example of what can happen when cognitive impairment and a deceptive online relationship intersect.

It also shows how artificial intelligence is changing the problem.

Romance scammers have always tried to make fake relationships feel real. AI can now help them create more convincing identities, photographs, voices, videos, and conversations.

And, as Bue's case demonstrates, sometimes AI can become the relationship itself.

For years, one of the most important questions in romance scam investigations has been:

Is the person you're talking to really the person they claim to be?

Now we need to ask another:

Is there actually a person there at all?

Sources and Further Reading


 
 
 

Comments

bottom of page