There is a lot of talk about how dangerous AI tools can be if they end up in the wrong hands. After all, for some time now we have been frightened by the phenomenon known as… deep fake, which is expected to become almost indistinguishable by humans over time.
The same applies to all kinds of methods of impersonating famous people in order to extort money. While recognizing fraudsters is relatively easy for the younger generation, older people are particularly vulnerable to such methods. Of course, that doesn’t sound very encouraging, does it?
What if we reversed the situation and used the potential of this technology to defend against fraudsters? This is exactly what the British telecommunications network O2 has undertaken. Effect? dAIsy, a special bot that is supposed to be a killer of all types of telephone fraudsters.
dAIsy – a chatbot that will waste the time of fraudsters
dAisy is a chatbot styled as a funny pensioner, which has only one goal – to waste the time of all kinds of telephone scammers. The choice of style is not accidental here, because it is older people who most often fall victim to this type of fraud. They are the ones the hero of today’s text is supposed to protect.
By dragging out conversations with scammers, the robot is supposed to make them have less time to try to cheat real people. In the promotional video you can hear that dAIsy performs its tasks perfectly. Some of the “victims” were losing their nerves, while the robotic pensioner didn’t care about it.
Interestingly, the funny chatbot is even able to provide bank details during a conversation. This is, of course, what such scammers are demanding. The trick, however, is that these are always completely false sets of information that provide no benefits to scammers.
It is worth noting here how the heroine of today’s text works. According to O2, during the conversation, the bot listens and converts the scammers’ speech into text and, using large language models, is able to generate a response that may resemble what a pensioner would say.
All this, of course, happens in real time and does not require any human intervention. Human participation was required only at the model training stage. Here, the help of legendary killers of telephone scammers, such as Jim Browing and Kitbog, was used.
Does dAIsy even work?
Representatives of the O2 network claim that their chatbot is effective and actually deals with fraudsters. According to dAIsy’s declaration, he is able to talk to fraudsters for up to 40 minutes, which in itself is an impressive result for a very “helpless retiree”.
The costs that O2 has incurred and continues to incur in connection with the development and maintenance of a large language model are unknown. Such things are usually not cheap and require a team of specialists who will ensure that everything works as intended.
However, it is certain that the company gains a lot from this in terms of image. After all, many people feel safer thanks to this and are able to trust such a company more. Of course, all kinds of fraudsters and scammers will not be happy, but no one asks them for their opinion.
