Pseudo AI: Humans Masquerading as Chatbots
Due to the high-cost and difficulty of developing and implementing AI-driven services some tech firms are resorting to pseudo AI workforces.
Replacing humans with Artificial Intelligence (AI) is not as easy as sci-fi movies would have you believe. In a volte-face, chatbots and automated AI systems are being quietly replaced or supplemented by human beings, while the technology catches up.
According to The Wall Street Journal (WSJ), companies find it easier and cheaper to hire a human workforce rather than build an AI workforce from the ground up. In some instances, companies are telling investors they have developed scalable AI technology while secretly relying on human intelligence.
Pseudo AI workforces are on the rise, and Gregory Koberger, CEO of ReadME, attributes this increase to the high level of technical and development challenges faced while implementing the technology. Koberger explained: “Using a human to do the job lets you skip over a load of technical and business development challenges. It doesn’t scale, obviously, but it allows you to build something and skip the hard part early on.”
Faux AI Raises Serious Privacy Concerns
The WSJ reported that many third-party app developers depend on humans and not AI when sifting through the personal messages of users. Edison Software claimed it used AI engineers to scan personal emails to improve its smart reply feature, however, it was, in fact, humans doing the legwork. Similarly, in 2008 Spinvox said it used AI technology to convert voicemails into text messages when in reality it had humans employed in overseas call centres to take on the laborious task.
Both these incidents highlight the potential for a data security failure, and while users may be comfortable with an AI bot trawling their sensitive data they are unlikely to feel safe knowing a human is looking at their private content.
Chatbot Workers Welcome Real AI Replacements
In 2016, Bloomberg exposed that chatbot scheduling services such as X.ai and Clara were subjecting their human workforce to spending 12 hours a day pretending to be chatbots. Workers said the job was so repetitive and dull that they were elated at the prospect of being replaced by real bots. In 2017, Expensify, a business expense management app, admitted it was using humans to transcribe some of their receipts although it had previously claimed that it had been using smartscan technology to process receipts.
Even a huge company like Facebook, with its vast pools of cash, has been using pseudo-AI. Their virtual assistant for Facebook Messenger also relied heavily on human workers despite the company’s heavy investment in AI. Some feel that this practice is dishonest and highly deceptive. Not only are the human workers being hidden away and marginalised but the customer is also being lied to. There is a risk that the lack of transparency regarding AI could contribute to the general mistrust around this emerging technology.