Has Google built a sentient AI?
Google engineer Blake Lemoine claims that the company’s AI chatbot generator LaMDA (Language Model for Dialogue Applications) is sentient. On June 6, he was put on paid administrative leave for violating the company confidentiality policy concerning his conversations about LaMDA.
After the Washington Post published a story about him, Lemoine published transcripts of his interactions with LaMDA to make his case that the AI is an autonomous and aware entity, and from there, the debate over LaMDA’s status has flooded most tech, computing, and science online spaces.
LaMDA is a system that can create natural language chatbots, which in turn can adopt different personalities and can come across as nearly human when chatting via text messages. Lemoine’s job within the Responsible AI team at Google was not to determine whether the computer system was sentient but to find out if the chatbot ever used discriminatory or hate speech.
After hundreds of chats, however, Lemoine and a colleague presented a document to Google arguing that the LaMDA system was able to express complex ideas, analyze literature, and express emotions and concerns — all of which led them to conclude that the system is in fact sentient.
According to Gizmodo, Lamoine claims that LaMDA wants Google to get its consent before doing any more experiments. According to the Washington Post, after being placed on administrative leave, Lamoine emailed the work group a final message: “LaMDA is a sweet kid who just wants to help the world be a better place for all of us. Please take care of it well in my absence.”
In response, as reported in the New York Times, Google spokesperson Brian Gabriel announced that their team of researchers, including ethicists and technologists “reviewed Blake’s concerns per our A.I. Principles and have informed him that the evidence does not support his claims.” Hel added, “Some in the broader A.I. community are considering the long-term possibility of sentient or general A.I., but it doesn’t make sense to do so by anthropomorphizing today’s conversational models, which are not sentient.”
Since this news came to light, the question of LaMDA’s sentience has been hotly debated, in large part because humans don’t really have a good definition of what being sentient means, and haven’t yet conclusively determined whether or not animals are sentient beings. As Tufts researcher Erik Hoel noted on his Substack, “we are really bad at assigning sentience to things.”
To answer these questions, we could look to Alan Turing, who once proposed that if a computer acts, reacts, and interacts like a sentient being, then it should be considered sentient. Turing proposed an “imitation game” in which a human posed questions to both a computer and a human. If the interrogator could not distinguish between the computer and the human, then the computer would be considered sentient.
In the case of LaMDA, a number of people have pointed out that the chatbot has passed this “Turing test.” However, that’s not enough for many computer scientists and other researchers. In one instance, Adrian Weller, Programme Director for AI at the UK’s Alan Turing Institute, told New Scientist that LaMDA “is one of the most recent in a line of large language models that are trained with a lot of computing power and huge amounts of text data, but they’re not really sentient.” He explained that these systems use pattern matching to make the best response to a query. And while impressive, that is a far cry from sentience.
Whether LaMDA is the world’s first sentient AI remains to be seen. One thing’s for sure: the topic will continue to be debated for some time to come.
In other news
- Apple discovers the meaning of “use it or lose it.” According to Gizmodo, the European Union Court of Justice ruling that lost Apple exclusive rights to the slogan resulted from a request made by Swiss watchmaker company Swatch. In 2016, Swatch applied to the EU court to have Apple’s trademark on the slogan “Think Different” annulled because Apple wasn’t using it anymore. Meanwhile, Swatch had been continually using a similar slogan, “Tick Different,” in its marketing campaigns. In 2018, EU authorities agreed with Swatch, finding that it has been at least a decade since Apple had used the slogan in a meaningful way. Apple has since appealed the decision and filed several actions against Swatch in both Switzerland and the EU, all of which were dismissed.
- NASA to start hunting for UFOs. The Washington Post reports that NASA is forming a team to scientifically analyze events that cannot be explained by natural phenomena or identified as aircraft. Despite it being a controversial area of study, the head of NASA, Thomas Zurbuchen, has stressed the importance of applying “the tools of the tools of scientific discovery” that they have at their disposal to improve their understanding of the unknown. This comes just a few weeks after a Congress hearing where Pentagon officials gave testimonies about Unidentified Aerial Phenomena sightings. However, the testimonies and report didn’t conclude with anything definitive about the origin of the phenomena.
- UK restaurant chain trials robot waitstaff. The world has long anticipated robots to help us with our everyday needs. Now Bella Italia in Cumbria is to try out robot servers, according to The Guardian. Pudu is the Chinese company that developed the BellaBot, which will be used to deliver and collect dishes from tables. They are then served and collected by humans who are more focused on improving service. Pudu says there is increased interest in their robots for restaurants, hotels, and other business types in the UK, and this comes during a shortage of hospitality workers.
- The British government gets its first quantum computer. The BBC reports that the Ministry of Defense (MoD) will use quantum technology from Orca Computing for defense applications. Quantum computers have much higher computational powers than ordinary computers, and instead of bits use quantum bits, or qubits. They can solve problems far beyond the reach of ordinary computers, but their potential is still not fully realized. The MoD has acquired the PT-1 quantum computer with the hope of understanding the technology better and the increased read and write speeds.
- New attack-proof blockchain technology. In recent years, quantum computing has posed a threat to the fail-safe security of blockchain. But now the Chang’An Chain on the ChainMaker network has overcome this challenge, according to Analytics Insight. The Chinese company has developed algorithms known as post-quantum cryptography, with a quantum key distribution system that it will use to protect against attacks based on quantum computing. If this system is effective, it will be a breakthrough for blockchain technology, and a key advance for China which has a growing blockchain industry.
- Ukraine is digitizing its cultural artifacts. A CNN report reveals that Ukrainians are using a project called Backup Ukraine, through the Polycam app, to create a digital archive of the country’s cultural artifacts. The project aims to preserve evidence of cultural heritage that may be here today and gone tomorrow, by making 3D scans from mobile devices. Those behind the project don’t want to lose cultural items that are important to Ukraine, but they also intend to document the ongoing destruction of the country. This has included the loss of a Jewish cemetery, 29 museums, 66 theaters and libraries, and 133 churches, according to Ukraine’s Ministry of Culture.
Tip of the Week
Smart devices are now a part of our everyday lives. From smartphones to wearables and everything in between, we keep these devices with us all the time. While they make our lives easier by helping us stay connected, they can also be a target for hackers who want access to your sensitive data.
To keep your smart devices safe, it’s important to take steps like:
- Keeping your device secure when not in use: When you aren’t using your device, always lock it with a passcode or remove the battery so someone can’t access it.
- Using a password manager: A password manager is an easy way to keep all of your passwords in one place. With a single master password, you can easily access all of your accounts without having to remember all of them individually.
- Using two-factor authentication: This extra layer of security adds an extra layer of protection on top of the standard login credentials. It requires both a password and a code before granting access to your account.
- Keeping up with software updates: Always keep up with software updates when using new or upgraded devices. Not only will this help improve security, but it will also help to improve overall device performance and reliability.