Google will warn people when they talk to its bot

Google Duplex bot, which mimics human speech, will include disclosure function

Tags: Artifical intelligenceChatbotGoogle IncorporatedUSA
  • E-Mail
Google will warn people when they talk to its bot Google's Duplex will have built-in features to identify itself as a bot in future, the company said.
By  Mark Sutton Published  May 12, 2018

Google will warn people if they are talking to its new bot, after a demonstration caused controversy.

The company demonstrated the new Duplex bot at its Google I/O Developer Conference last week.

Duplex is able to mimic human speech with pauses, slang and other tics of speech to make it sound more convincing. Google said it has developed the bot to carry out real world tasks over the phone

In the I/O demo, the bot was used to book a hairdressing appointment and enquire about a restaurant reservation. In both cases, the human on the other end of the call didn't appear to realise they were talking to an artificial voice.

The bot uses Google's DeepMind WaveNet software which has analysed conversations and can mimic them in ‘natural' conversations.

The demonstration caused outrage from some sectors of the press, with Duplex (and Google) variously described as "terrifying", "deceptive" and "unethical".

Google has said that it is still fine-tuning the system, and will make sure that it does identify itself in future.

"We are designing this feature with disclosure built-in, and we'll make sure the system is appropriately identified," the company said in a statement.

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code