MONROE, Conn. — Changlin Lee, executive director of the AI Safety Awareness Project, appeared at the Monroe Senior Center on June 16 to speak on the topic of AI, but before speaking about the potential problems and threats posed by the emerging technology, she demonstrated some of its benefits.
Before typing questions into ChatGPT, Li told center users to think of artificial intelligence (AI) like an overzealous intern. He then used an app on his phone to talk to ChatGPT instead of typing.
“Hello, ChatGPT, can you hear me?” Ms. Lee asked on speakerphone.
“Yes, Changlin, I can hear you,” the bot responded in a friendly male voice. “what’s on your mind?”
Lee said he is planning a trip from Monroe to the Bahamas. The chatbot said it would confirm the flight. I said, “Do you want something relaxing? Let’s plan a snorkeling trip.”
Lee said he hopes to fly out of Westchester Airport on July 31st. The bot told them they would likely have a connecting flight in Miami and needed to arrive at their destination by August 1st, adding: “I recommend a short trip to the beach.” “Easy and refreshing.”
“It’s very similar to how we interact with humans,” Lee said. “We might run into some problems here.”
Criminals often use AI to imitate the voices of Mark’s relatives and friends to trick them into sending money. During the demonstration, Lee showed an AI-generated audio clip using an audio recording of the reporter, Monroe Sun editor Bill Bitter. They included angry voicemails and recordings in which Bitter, as a parent, called her child’s school and told corporate headquarters to have a “family friend” come pick up her daughter.
Lee also had the AI generate a fake photo using two family photos that Vitale submitted for demonstration. In one AI-generated photo, Vitale is seen sitting in a cell wearing an orange prison jumpsuit, and in another photo he is seen wearing camouflage and holding a hunting rifle.
Imperfections can sometimes be found in AI-generated photos, Lee said, so online criminals could generate slightly out-of-focus photos to hide details. He showed what appeared to be a black-and-white surveillance camera screenshot of Vitale standing near a woman in an elevator wearing a hoodie and looking intimidating.
Lee also used AI to create a realistic New York Times article that showed the Monroe Sun’s editor-in-chief was embroiled in a scandal. These tactics have been used against politicians on both sides of the aisle during presidential elections.
Lee also showed how, in a targeted attack, the AI could create someone’s identity with a backstory and generate a convincing phishing email to Bittar’s LinkedIn profile to steal his password.
On June 16, Monroe Police Department Kyle Stevens participated in a presentation along with Officer Danny Leon. Co-director Madeline McBride and facilitator Lauren Mecca from the AI Safety Awareness Project also attended.
The AI Safety Awareness Project is a 501c3 nonprofit organization whose mission is to ensure everyone has knowledge, opinion, and voice in the direction of AI as a technology.
“We want to take the first step,” Lee said. “We don’t want to stay at the second step. It’s up to you. Talk to your family and friends. What excites you most about AI? What do you want the future to be?”
While AI can do some great things, Lee said his nonprofit is focused on risk because it is a safety organization. The AI Safety Awareness Project provides presentations at police departments, libraries, city and state agencies, and community centers.
What is AI?
Just as “overeager interns” sometimes make mistakes, Lee said, AI can also make mistakes. In that case, AI is different from software and cannot check lines of code for errors, he said. AI knowledge grows by learning from large amounts of data. Lee suggests fact-checking AI the same way you would an intern, checking different sources to verify information.
AI is trained, not designed, Lee said. “Even creators don’t know how they create their work,” he says. “It’s more of a process of growing something than engineering something. There’s no universal official definition of AI.”
To give a working definition, he defined it as a machine that imitates aspects of human intelligence.
Lee said the first form of AI was a thermostat in the 1940s, followed by Deep Blue defeating a chess champion in 1997. However, it was only programmed to play chess. In 2011, Watson won “Jeopardy!” And in 2016, AlphaGo defeated Go Champ. DALLE Image Generator arrived in 2021, and ChatGPT exploded in functionality in 2022.
Lee said there are AI large-scale language models (LLMs), systems that can be trained on vast amounts of text to understand and generate human-like language by modeling the next word.
There are also generative AI systems that emerged around 2022. He said it creates new content (text, images, videos) from existing data and generates original output based on prompts.
“They write songs for you,” Lee said.
To illustrate how far AI has come in a short period of time, Lee showed an AI-generated video of actor Will Smith eating spaghetti. Smith’s face looked good in the video, but the way he sloppily shoved spaghetti into his mouth looked a little goofy.
“This is terrible, isn’t it?” Lee asked. “This is the best thing we can do in March 2023.”
He then showed a generated video of a police traffic stop that took place in August 2025. Clear video showed the officer walking next to the car. “You scored 90 points in a 40 point zone!” he barked at the driver.
“I’m sorry, but could you please smooth things over?” the driver asked, handing the officer a wad of cash.
The officer took the money and said, “Get out of here before I…” The audience at the senior center laughed at the scene.
“This is a completely fake video of a road stop,” Lee said, adding that there was no driver, police officer or car present. “It was completely AI-generated,” he says. “In 36 months, I went from being laughably bad to being like, ‘Hey, this is really real!'”
Lee said the speed of AI is also rapidly accelerating, with task completion doubling every seven months and completing every four to five months.
Problems that AI can cause
Lee shared three pillars of concern about AI.
- Will it exacerbate existing problems such as fraud and fraud, concentration of power and isolation?
- Will jobs such as travel agency become obsolete? What does that mean for labor? Should we move to a world where things are replaced by AI? If so, what should it look like?
- loss of control. The people building the system don’t understand why the system behaves the way it does. If we don’t know why the system behaves the way it does, how can we prevent it from “cheating” and killing everyone?
Resources are another issue. Data centers not only use water to cool their systems, they also use a lot of energy for AI.
During the question and answer period following Lee’s presentation, Detective Stevens told the audience how AI is making it harder for police to investigate fraud cases. He encouraged residents to contact police whenever they have questions about possible fraud.
Seniors expressed concerns about the growing adoption of AI, and Lee joined the Aisafetyawarenessproject.org mailing list to let members of Congress know that AI safety is an important issue to them and encourage them to get involved by educating themselves. He also shared other resources such as BlueDot Impact.
“I’m optimistic and idealistic about this,” Lee said. “You don’t need a billion dollars to talk to your legislators.”
Respectful comments that include the commenter’s first and last name are welcome.
