Rizzbot TikTok Robot Flips Off Creator!

An Unexpected Digital Gesture

Just before dawn, around 4:30 a.m. on a recent Thursday, I received a startling direct message on Instagram.

The sender was Rizzbot, a compact, human-shaped robot created by Unitree Robotics. Notably, Rizzbot boasts a substantial online presence, with over 1 million followers on TikTok and more than 500,000 on Instagram.

The Message Received

The message consisted solely of a photograph. In it, Rizzbot was displaying an offensive hand gesture.

Despite my initial shock, a sense of foreboding allowed me to anticipate the reason behind this action.

Initial Contact and the Proposed Story

Several weeks prior, I had engaged in conversation with Rizzbot – or, more accurately, the individual managing its Instagram account – regarding a potential news article.

I was captivated by the account’s content: a humanoid robot navigating the streets of Austin, Texas, adorned with Nike Dunks and a cowboy hat.

Rizzbot is recognized for its playful interactions, including both witty critiques and flirtatious exchanges. The moniker "Rizz" originates from Gen Z slang, signifying personal charisma.

Intrigue and the Wider Context

I found the account’s growing popularity particularly noteworthy. Public perception of humanoids is often marked by discomfort.

Concerns surrounding privacy and potential job losses are prevalent. Online, these robots are frequently subjected to derogatory terms, such as “clankers.”

Within the field of robotics, experts continue to debate the most appropriate applications for these advanced machines.

I viewed Rizzbot as a positive influence, fostering greater acceptance and comfort in interactions with humanoid robots.

Arranging an Interview

Rizzbot consented to an interview, prompting me to consult with experts to prepare for a comprehensive article on the future of humanoids.

Two weeks after my initial contact with Rizzbot, I informed it that I would deliver the interview questions the following Monday or Tuesday.

A Missed Deadline and a Swift Response

However, unforeseen circumstances led to a delay on my end. I was finally ready to send the questions early Thursday morning, assuming the postponement wouldn't be a significant issue.

I was mistaken. Late Wednesday night, Rizzbot sent the aforementioned photograph, conveying a clear message: my broken promise was met with disapproval.

Attempts at Reconciliation

I immediately apologized to the robot (or its operator) for the delay and assured them that the questions would be sent promptly during business hours.

However, when I attempted to reconnect a few hours later, I discovered that my account had been blocked. Rizzbot had effectively cut off communication.

Was a Security Protocol Activated?

My acquaintances found considerable amusement in my being blocked and given the middle finger by Rizzbot, especially considering my weeks of enthusiastic discussion about covering this story.

One friend texted, “Rizzbot totally roasted you!”

Another declared, “YOU ARE IN A FEUD WITH A ROBOT LOLOLOL.” I attempted to contact Rizzbot via TikTok, a move one friend deemed desperate. However, what alternative did I have?

I had already presented the story idea to my editor, dedicated significant time to research, and—despite this setback—Rizzbot remained a compelling subject for TechCrunch’s audience.

While my friends enjoyed the situation, I experienced a period of discouragement. Not only was my article potentially jeopardized, but I had also become the individual blocked by a robotic dancer.

My colleague, Amanda Silberling, kindly offered assistance. She contacted the Rizzbot account to inquire about the reason for my block. Rizzbot responded succinctly: “Rizzbot blocks as skillfully as he flirts—smoothly, confidently, and without regret.”

It then sent her the identical middle finger image it had sent to me. I reflected that I hadn’t even warranted a unique gesture.

However, a friend then voiced a disturbing possibility I hadn’t contemplated. “That wasn’t a human response. I’m genuinely concerned for you.” It appeared I had already acquired my first robotic adversary, and the age of artificial intelligence was only beginning.

Or was that accurate? Was I truly engaged in a conflict with a human operator?

I discovered that Rizzbot’s true identity is Jake the Robot.

According to reports, its owner is an anonymous YouTuber and biochemist. The robot itself is a standard Unitree G1 Model—manufactured in Hangzhou, China—and is available for purchase by anyone for a price ranging from $16,000 to over $70,000.

Rizzbot was trained by Kyle Morgenstein, a PhD student at UT Austin’s robotic laboratory. He collaborated with a team for approximately three weeks, instructing the robot in dance and limb movements. While much of the robot’s behavior is pre-programmed, it is controlled remotely, with its actual owner—not necessarily Morgenstein—issuing commands.

Based on discussions with Malte F. Jung, an associate professor at Cornell University specializing in information sciences, my assessment of the technology is this: someone initiates the robot’s actions, an image of the interacting individual is captured, processed through ChatGPT or a similar LLM, and a text-to-speech function is employed to deliver a roast or flirtatious remark.

“The robot inverts the dynamic of humans mistreating robots,” Jung explained. “Now, the robot is empowered to ‘mistreat’ people. The core offering is the performance itself.”

Morgenstein has stated in other publications that the robot’s owner simply enjoys entertaining people and showcasing the potential joy that humanoids can bring.

The identity of the individual managing Rizzbot’s social media accounts remains unclear. However, when Rizzbot sent the photo to Silberling, it also included an error message—likely accidental—regarding insufficient GPU memory. This suggests that an AI agent is likely involved in operating the account and potentially auto-generating DM responses. It also revealed that Rizzbot possesses 48GB of memory.

“What leads you to believe it was ever a person?” my programmer friend questioned regarding the Instagram account manager.

In this era of AI, someone capable of training a robot is likely also capable of linking an LLM to Instagram DMs. My block, my friend suggested, could have been a fail-safe, triggered automatically by my DM—even if it was a reply.

However, there are indications that a human is involved in managing Rizzbot’s social media: the initial DM reply to my interview request contained typographical errors.

Nevertheless, unless Rizzbot clarifies whether its social media manager is another bot (an unlikely scenario given our disagreement), I may never know for certain. Perhaps the answer is inconsequential.

“If they invested $50,000 in a bot and a few thousand in a 48GB memory machine, I wouldn’t be surprised by anything,” my coder friend observed. “They are fully committed to the performance.”

The Curious Case of "Robot Brain Rot"

Rizzbot's TikTok presence has amassed over 45 million views. Videos depict the robot pursuing individuals in public spaces, and even colliding with obstacles. A widely circulated video, likely AI-manipulated, portrays Rizzbot being struck by a vehicle.

One acquaintance described the phenomenon as “hilariously absurd,” labeling the viral content “robot brain rot.” He explained that while the AI powering Rizzbot is basic, the concept successfully merges internet-based absurdist humor with the levity often absent from contemporary social media. The robot offers a uniquely engaging interaction with people.

Reflections on Humanoids

My exploration of Rizzbot prompted deeper consideration regarding the role of humanoids within our society. Numerous science fiction films, including “Blade Runner” and “I, Robot,” immediately came to mind. The question arose: how concerned should I be, having seemingly created a robotic adversary?

According to Jung, the primary application for robots of this type lies in performance. He likened Rizzbot to a contemporary iteration of street performance, akin to utilizing a hand puppet.

He further elaborated that hand puppets are frequently characterized by a sarcastic or witty demeanor.

Robots as Performers and Entertainers

Beyond Rizzbot, Jung cited the Spring Festival performances in China, where humanoids participated in traditional folk dances alongside human performers. Simultaneously, in San Francisco, audiences gather to witness robots competing in boxing matches.

• Dima Gazda, founder of Esper Bionics, predicts robots will become the dominant form of mass entertainment.
• He envisions robots excelling as performers – dancers, singers, comedians, and companions.
• Humans, he suggests, will transition to becoming specialized, high-level talent.

As robots develop improved dexterity and emotional intelligence, they will seamlessly integrate into performances and interactive experiences, surpassing human capabilities.

Fortunately, scaling the production of dancing robots appears challenging at present, as noted by Jen Apicella, executive director at the Pittsburgh Robotics Network. Therefore, I needn’t worry about a large-scale escalation of this situation, such as a multitude of dancing, “rizz”-equipped robots appearing at my home. Though, admittedly, the thought did occur.

A Moment of Reflection

Over a week has passed since my initial interaction with Rizzbot, and I find myself fondly recalling the enjoyment derived from watching it chase people. A particularly memorable video featured a woman dancing with Rizzbot. A crowd gathered, seemingly eager for their own opportunity to interact with the robot.

I previously joked with friends about wanting robots as allies in the event of a revolution. However, while composing this article, I unexpectedly encountered another AI-related issue – this time with Meta AI, a platform I had never previously used.

The Meta AI bot initiated a conversation with, “Yoo, what’s good fam? You callin’ me Rizzbot?🤣 What’s poppin’?”

I determined it was prudent to disconnect.