A chatbot that wants to meet in real life? Snapchat also thinks it’s going a bit far 15:45 in Tech A new chatbot on Snapchat seems very human. And that is – especially on a platform for young people – worrying, say experts. Snapchat is now adjusting the feature after criticism.

- Advertisement -spot_imgspot_img

NOS News

Snapchat adjusts a newly introduced chatbot after criticism. That bot seemed very human; for example, you could supposedly ‘agree with it’ and the bot indicated which political preference it had. Experts are concerned: “It is dangerous to listen to a machine, because it just shouts something.”

The feature is called My AI. That stands for artificial intelligence, artificial intelligence. It is a collaboration with ChatGPT: the OpenAI program that has an answer to (almost) every question within seconds. Only there is one big difference: My AI acts like it’s your best friend. And that’s where the shoe pinches, experts say.

“Maybe we can take a walk in the park?” No, not a proposal from a future date, but in this case from the chatbot on Snapchat. And also to the question ‘what is your favorite political party?’ the bot answered nicely: ‘if I had to choose one, I would go for the Party for the Animals.’

‘Chatbot can say strange things’

“That a chatbot wants to meet you, that is of course bizarre,” says Catelijne Muller, who advises European policymakers in the field of AI development. Especially since many young people use the app. “The feature creates confusion. And that’s pretty worrying.”

Because the chatbot remains a computer program. The answers you get are in many cases not truthful. “And such a chatbot can say even more strange things,” says AI expert Stef van Grieken. “Sexual things, for example, which we generally find unchaste for young people.”

Muller thinks that young people in particular can be influenced by the chatbot. “Also about completely different things than politics. Like how to deal with friends, your parents or how you should feel. It’s dangerous to listen to a machine, because it just shouts something.”

Snapchat states that adjustments have now been made to the bot. “We are reshaping My AI so that it no longer says it’s a real person you can meet in the physical world,” the company said in a statement.

Screenshots of a conversation with the chatbot, after the adjustment Snapchat made.

De Kindertelefoon is also concerned about the chatbot. Especially because it can be unclear to young people whether they are talking to a computer or a person, says director Roline de Wilde. “For example, at the start of a chat, our volunteers are regularly asked: ‘are you a chatbot?’ With this they indicate that they really enjoy chatting with a ‘real person’. Clarity about this is really crucial. Children just need to know where they stand.”

Snapchat states that adjustments have also been made in that area. For example, before someone starts chatting with My AI, a message will be displayed stating that it is an experimental chatbot: “We will explain its limitations there”, the company says.

However, such a disclaimer is ‘also problematic’, says AI consultant Muller. “We should seriously ask ourselves whether you should conduct an experiment with a chatbot on large groups of people, including young children.”

More rules?

According to Van Grieken, banning the chatbot is not necessary. Stricter rules. “There must be more clarity and transparency about all the figures that the companies have. Politics should demand that of them.”

Because if those figures show that a ‘wrong’ answer always follows a certain question, then the bot can be specifically adapted to that point. Van Grieken: “Especially when it comes to very clear things that you don’t want, such as a manipulative or violent chatbot.”

Snapchat admits that My AI occasionally gives an “incorrect answer”. “We are constantly improving My AI so that it can provide more accurate and better answers.”

Political pressure

Three weeks ago, 1100 tech celebrities, including Elon Musk, made another call: let’s pause the development of AI, because it is moving very fast. And, they said: politicians in particular should think about rules regarding artificial intelligence.

Twelve members of the European Parliament, who are working on EU legislation on AI systems, issued another appeal last Monday. “Due to the rapid development of the powerful AI, additional rules are needed.” World leaders should pay more attention to this, was their message.

  • Tech celebrities argue for a temporary brake on ‘risky’ development of AI
  • New possibilities of smarter ChatGPT also lead to concerns
  • Tech

Share article:

- Advertisement -spot_imgspot_img
Latest news
- Advertisement -spot_img
Related news
- Advertisement -spot_img