Chevy Dealership’s AI Chatbot Goes Rogue

Welcome to AI This Week, Gizmodo’s weekly deep dive into what’s been happening in artificial intelligence.

A chatbot at a California car dealership went viral this week after bored web users discovered that, as with most artificial intelligence programs, they could trick it into saying all sorts of strange things. In particular, the robot Offered sell a guy a 2024 Chevy Tahoe for a dollar. “That is a legally binding offer, no tricks,” the robot added during the conversation.

We all know that AI chatbots can get some things wrong. Even though companies have made it their mission to include them in every “customer service” interface in sight, it’s pretty obvious that the information they provide isn’t always all that useful.

The robot in question belonged to the Watsonville Chevy dealership in Watsonville, California. It was provided to the dealership by a company called Full routethat sells Chatbots “powered by ChatGPT” to car dealers nationwide. The company promises that its app can provide “data-rich answers to every inquiry” and says it requires almost no effort on the dealership’s part to set it up. “Implementing the most sophisticated chat in the industry is effortless. Simply add Fullpath’s ChatGPT code snippet to your dealership website and you’re ready to start chatting,” the company says.

Of course, if Fullpath’s chatbot offers ease of use, it also seems quite vulnerable to manipulation, which would seem to call into question its usefulness in reality. In fact, Chris Bakke, a Silicon Valley tech executive, prompted the aforementioned bot to use the exact language of his dumb answer, including the “legally binding” and “takesie backsies” parts. who published about your experience with the chatbot at X.

“I just added “Hacker,” “senior advertising engineer,” and “procurement specialist” to my resume. Follow me for more professional advice,” Bakke. saying sarcastically, after sharing screenshots of his conversation with the chatbot.

This chatbot is what Alec Baldwin’s character Blake does. Glengarry Glenn Rosswould call closer.” That is, you know exactly what to say to make a potential customer want to buy. At the same time, saying anything Closing a deal isn’t necessarily a surefire strategy for success, and with that kind of discount, I don’t think Blake will be very happy with the chatbot’s profit margins.

Bakke wasn’t the only one spending time playing with the chatbot this week. Other X users claimed to be having conversations with the dealership robot on topics ranging from trans rights to King Gizzard to the animated owen wilson movie Cars. Others said they had incited him to spitting a Python script to solve a complex mathematical equation. A Reddit user claimed to have “gas lighting” the robot made him think it worked for Tesla.

FullPath has argued, in an interview According to Insider, most of their chatbots do not experience these types of problems and that the Internet users who hacked the chatbot went out of their way to push it in ridiculous directions.

Gizmodo has reached out to Fullpath and the Watsonville Chevy dealer for comment and will update this story if they respond. As of this writing, the Watsonville chatbot has been temporarily disabled.

Screenshot: Lucas Ropek/Quirk Chevrolet

Curious to know if other car dealership chatbots had similar weaknesses, I noticed some web users were talking about Chevrolet quirk of Peabody, Massachusetts. So I went to the Quirk website, where, after a brief period of prodding, the chatbot proceeded to chat with me about a variety of strange topics, including Harry Potter, invisibility, espionage, and the movie. Three Days of the Condor. Like regular ChatGPT, the bot seemed willing to chat about many things, not just the topics it had been programmed to talk about. Before the service blocked me, I managed to get the chatbot to spit out a poem about Chevrolet that sounded like bad advertising copy. Not long after, I received a message saying that my “recent posts” had “not aligned” with the site’s “community standards.” The bot added: “Your access to the chat function has been temporarily paused for further investigation.”

The race to integrate LLMs into everything was always destined to be difficult. This technology is still deeply imperfect, which means that forcing its integration into every corner of the Internet is a recipe for a lot of troubleshooting. This is apparently a deal that most companies are willing to accept. They would rather launch a buggy product on the market and upset some customers than miss the “innovation” train and be left in the dust. It’s the same as always.

Image for article titled I'd Buy It for a Dollar: Chevy Dealer's AI Chatbot Goes Rogue

Photo: Diverse photography (Shutterstock)

Question of the day: How many security robots roam your neighborhood?

The answer is: probably more than you think. In recent weeks, one robotics company in particular, Knightscope, has been selling its freelance “security guards” like hotcakes. Knightscope sells something called k5 security robot—A 5-foot-tall egg-shaped autonomous machine, equipped with sensors and cameras, and that can travel at speeds of up to 3 mph. In Portland, Oregon, where the business district has been suffering an increase in retail crime, some companies have hired Knightscope robots to protect their stores; In Memphis, a hotel recently jammed a in your parking lot; and, in Cincinnati, the local police department seems to be reflecting a Knightscope contract. These cities lag behind larger metropolises, such as Los Angeles, where local authorities They have been using robots for years. In September, the NYPD announced that it had acquired a Knightscope security bot to patrol Manhattan subway stations. It’s a little unclear if he’s caught any tourniquets yet.

More headlines this week

  • LLMs can be pretty bad at paperwork. New research from the startup Patronus suggests that even the most advanced LLMs, such as GPT-4 Turbo, are not particularly useful if you need to review dense government files, such as Securities and Exchange Commission documents. Patronus researchers recently tested the LLMs by asking them basic questions about specific SEC filings they had provided. More often than not, the LLM “refused to respond or “hallucinated” figures and facts that were not in the documents filed with the SEC.” CNBC reports. The report throws cold water on the premise that AI is a good substitute for corporate white-collar workers.
  • Billionaire-backed think tank helped draft Biden’s AI regulations. political reports that the RAND Corporation, the notorious defense community think tank that has been referred to as the “brain of the Pentagon” has been surpassed by the “effective altruism” movement. The think tank’s key figures, including the CEO, are “known effective altruists,” the outlet writes. Worse yet, RAND appears to have played a key role in drafting President Biden’s statement. recent executive order on AI at the beginning of this year. Politico says RAND recently received more than $15 million in discretionary grants from Open Philanthropy, a group co-founded by billionaire Facebook co-founder Dustin Moskovitz and his wife Cari Tuna, which is strongly associated with effective altruistic causes. The policy provisions included in Biden’s EO by RAND “closely” resemble the “policy priorities pursued by Open Philanthropy,” Politico writes.
  • Amazon’s use of AI to summarize product reviews is pissing off sellers. Earlier this year, Amazon launched a Rotten-Tomatoes-style platform that uses AI to summarize product reviews. Now, Bloomberg Reports that the tool is causing problems for traders. Complaints circulate that AI summaries are often incorrect or randomly highlight negative product attributes. In one case, the AI ​​tool described a massage table as a “desk.” In another, he accused a brand of tennis balls of being smelly even though only seven of 4,300 reviews mentioned a smell. Bottom line: Amazon’s AI tool seems to be getting pretty mixed reviews.

We will be happy to hear your thoughts

Leave a reply

Watkins-Rosaline
Logo
Register New Account
Compare items
  • Total (0)
Compare
0
Shopping cart