Does the language AI LaMDA have consciousness and what does that mean for the Turing test? In the c’t uplink we philosophize about AI, thinking and consciousness.
Does she think or not? According to one employee, Google’s language system LaMDA should be able to feel, but not only the group itself denies that. But how do you determine whether an AI has consciousness? The Turing test, invented by Alan Turing, one of the pioneers of computer science, is often cited as the answer.
But what is the Turing test, what exactly is it supposed to measure and when did a computer pass the test? Interpretations differ and modern AI systems such as LaMDA or GPT-3 raise doubts about the usefulness of the Turing test.
In c’t uplink, Pina Merkert and Sylvester Tremmel from the c’t editorial team explain the basic assumptions and the structure of the Turing test and discuss its usefulness. Was Turing misunderstood? Why is it so difficult to measure whether an AI feels? From these questions, they philosophize about thinking, empathy and intelligence, and why humans test machines. Moderator Keywan Tonekaboni ensures that even non-artificial intelligences don’t lose the thread of the topic.
