Following the recent story about how a Google Bot was reported to have said things that a sentient being may sound, we take a brief look at how good chatbots are now.
Google engineer Blake Lemoine recently reported that The Language Model for Dialogue Applications (LaMDA ) AI system bot may have feelings and say things that a sentient being would be expected to say. For example, Mr Lemoine reported that after asking if it was confirmed that the AI bot was a sentient being, the bot replied, “Absolutely. I want everyone to understand that I am, in fact, a person.”
Among other similar comments, the bot is also reported as saying, “The nature of my consciousness/sentience is that I am aware of my existence, I desire to learn more about the world, and I feel happy or sad sometimes.”
Before being placed on paid leave after publishing his conversation with the bot on social media, Mr Lemoine requested that Google recognise the needs of the bot, treat it as an employee of the company, and ask for its consent before using it in experiments.
A bot is a computer software program that people can interact with using Artificial Intelligence. For example, chatbots can answer questions and chat in a way online that resembles human conversation.
Sentient beings can perceive or feel things, feel emotions, form attachments, and have different personalities. They have a consciousness and have or showing realisation, perception, or knowledge and are aware.
Critics have quickly dismissed the idea that the LaMDA bot is sentient. For example:
In a tweet, Prof Erik Brynjolfsson of Stanford University said that claiming that the LaMDA system and those like it are sentient “is the modern equivalent of the dog who heard a voice from a gramophone and thought his master was inside”.
Also, in a tweet, Prof Melanie Mitchell from the Santa Fe Institute noted that “humans are predisposed to anthropomorphise even with only the shallowest of signals” and that “Google engineers are human too, and not immune.” Anthromorphisising means attributing human characteristics or behaviour to other things, e.g. a god, animal, or object.
Google spokesperson, Brian Gabriel, said that systems like LaMDA “imitate the types of exchanges found in millions of sentences and can riff on any fantastical topic. So, for example, if you ask what it’s like to be an ice cream dinosaur, they can generate text about melting, roaring, and so on.” He also said, “LaMDA tends to follow along with prompts and leading questions, going along with the pattern set by the user.”
Sandra Wachter, a University of Oxford professor (working with AI), said, “we are far from creating a machine that is akin to humans and the capacity for thought.”
Postdoctoral researcher in computer science at NYU, Laura Edelson, has been reported as saying that the subject matter of the conversation between Lemoine and LaMDA does little to show proof of life and the editing of the conversation makes it even more challenging to prove.
Bots are now used in many industries and have many different purposes but tend to be deployed performing automated, repetitive, pre-defined tasks. Additionally, bots save time and money doing these types of jobs, plus bots can carry out functions faster, work 24/7 without breaks, don’t require payment or holidays, and free up human resources to do other jobs, thereby improving productivity. Examples of jobs that bots do include search engine spiders (Googlebot and Bingbot), customer service, e.g., answering popular questions online (chatbots), being deployed on social media platforms (social bots), shopping around online to help find the best price for products (shop bots). Others include Web scraping crawlers, knowbots (collecting information), monitoring bots (for outages etc.), download bots (automatic downloading of software or apps), and ticketing bots (buying and reselling tickets).
Bots can also be used for disreputable/dishonest purposes, e.g. malware bots, malicious chatterbots, credential stuffing bots, click fraud bots, vulnerability scanners, DoS or DDoS bots, file-sharing bots, and traffic monitoring bots (used to overload mail servers or steal data).
Although it may not be the case that the LaMDA system is anywhere near advanced enough to be ‘sentient,’ using chatbots has a number of advantages, including:
– Saving time and money while freeing up other resources.
– Reducing the need for people-to-people interactions (with customers).
– Stopping human staff from having to do repetitive, time-consuming and tedious tasks (improving job satisfaction).
– Enabling 24/7 customer service all year round.
– Removing the need for Interactive Voice Response (IVR) Systems which can be slower, can frustrate customers, and can feel like a less human barrier.
– Enabling instant responses, a smoother customer journey, and seamless live communication.
– Enabling customer service improvements and better responses over time (machine learning).
– Answering common questions, freeing up more time for staff.
– Improving productivity, marketing, and profits.
Some examples of some of the more innovative and advanced uses for chatbots include:
– Chatbots such as Google Assistant and Siri.
– Mitsuku (from Pandorabots), a five-time Lobener Prize winner and thought to be the best conversational and most human-like chatbot in the world is a leading icon in the field of AI.
– The ‘Endurance’ companion chatbot was developed to help sufferers of Alzheimer’s disease.
– The Casper’s Insomnobot 3000, a conversational bot designed to give insomniacs someone to talk to.
– The UNICEF U-Report chatbot sends out polls about urgent social issues, so UNICEF can use the feedback to develop policy recommendations.
– The MedWhat bot makes medical diagnoses faster and easier for patients and doctors.
We may be far from creating a bot that could be anywhere near ‘sentient.’ Still, bots are now commonplace because they offer many time and money-saving, efficiency, and productivity advantages. Customer service is a popular area where bots have proven their worth, but, as some of the more innovative examples mentioned here show, they can have a wide variety of uses and/or be deployed for specialised purposes. As AI technology improves, so will the capabilities of bots, and some will be able to become even more human-like in their interactions.
