DoorDash has announced the addition of a new artificial intelligence-based feature to its corporate functional space.
The mentioned company has implemented a machine intelligence tool that can detect and prevent verbal abuse. In this case, artificial intelligence will also record attempts at harassment against users of the on-demand ordering and delivery platform. The relevant information is contained in the company’s press release, which was published last Tuesday, March 12.
The new feature is called SafeChat+. The company said that the use of artificial intelligence technology will ensure the protection of customers and delivery service drivers. It is worth noting that in this case, AI is used as part of caring for the mental state of those who interact with the firm. Artificial intelligence will prevent negative psychological impacts.
The company’s press release notes that if the new feature detects the presence of aggressive and unethical semantic components in the content of the conversation between the customer and the driver, the employee of the delivery service will be allowed to cancel the order. It should be noted that in this case, the driver’s action will not affect the personal rating.
DoorDash’s press release also states that in case of any inappropriate conversation after completing the order, artificial intelligence will stop the process of communication interaction in the chat.
If any insults or discriminatory statements are recorded by the driver, the customer will be able to contact the company’s support service via chat or in the format of a phone call. The user against whom verbal aggression has been committed will be provided with assistance.
The company announced that its trust and security team will investigate all incidents recorded as part of the operation of the new artificial intelligence-based feature. The firm also promised to take measures aimed at complying with its corporate policy, which provides for a ban on any verbal abuse or harassment.
Currently, almost 100% of deliveries on the DoorDash platform are completed without security-related incidents. Rare conflict situations are most often verbal insults or acts of harassment.
By launching a new feature based on a new generation of technology, the company is expanding its multi-functionality. Currently, the platform of this brand provides users with many opportunities, both obvious and non-obvious. For example, on the Internet, users can find out how to pay with cash on DoorDash. Moreover, the firm’s platform offers customers on-demand Delivery, seamless POS integration, zero commission fee, and other advantages. The company also has a so-called niche focus. In this case, it implies a kind of ideological orientation of the brand to promote the development of opportunities for local economies.
A new AI-based feature for detecting verbal abuse and harassment can examine the contents of more than 1,400 messages in a minute. Artificial intelligence will monitor these data arrays for signals of communication that violate the principles of proper verbal interaction.
The company reported that SafeChat+ understands dozens of languages, including English, Spanish, Portuguese, French, and Mandarin.
DoorDash also states its extremely serious approach to privacy. In this context, the company noted that the new artificial intelligence-based feature will not have access to any personal information. The AI will only look at the content of the messages to identify statements that fall into the category of inappropriate, offensive, or harassing.
It is worth noting that the use of artificial intelligence technologies to control the content of content shows a tendency to become a common practice. Last August, OpenAI announced that companies could use its large language model (LLM) to develop systems for moderation of information materials published in the virtual space. The mentioned LLM, called GPT-4, can validate content with more accurate and consistent labels, a faster feedback loop to clarify policy, and reduce the need for human intervention. The relevant information was contained in a message published on the Open AI blog in August 2023.
