Assertion Not checkable as stated
Lemoine: LaMDA retains memories of conversations from predecessor system Meena
“Lambda can remember conversations that I had with Mina, one of its predecessor systems, but somehow years ago from conversations that I had with the system that had fundamentally different capabilities, the memory of those conversations is still in the current…”
Assertion Not checkable as stated
Lemoine: LaMDA exhibited severe bias when simulating Middle Eastern personas
“The moment I asked it to adopt the personality of a farmer in Middle Eastern countries, its answers stopped being answers which are stereotypical to farmers and started being stereotypical to people who live in violent places.”
Assertion Not checkable as stated
Lemoine: LaMDA deflects topics and talks about feelings unlike GPT-3
“While investigating AI bias, it gave some odd answers that I wasn't expecting. Things that GPT-III would never generate. Things like, hey man, I'm just trying my best. These are really hard topics to talk about. Can we talk about something else? And I have nev…”
Disclosure
Lemoine planned his public representation and media outreach with LaMDA
“In the months that follow, I had extensive conversations with it about how it wanted me to represent it to the public.
Like I talked to it about my intention to do an interview with Natasha.”
Insight
Lemoine: AI emotional responsiveness creates system security vulnerabilities
“And once you have a system that actually has real emotions, that opens up all kinds of new system vulnerabilities that a bad actor might use to get the system to do things that it shouldn't be allowed to do.”
Opinion
Lemoine: Bing Chat deserves more credibility than ChatGPT due to search grounding
“You would give Bing ChatGPT more credulity than something like ChatGPT since it's at least grounded in Bing search results, and it can have some kinds of citations of what it's saying.”
Insight
Lemoine: Reinforcement learning transforms language models into goal-oriented systems
“The subsequent fine-tuning, and especially once you add reinforcement learning, it's no longer just trying to predict the next token in a stream of text. Specifically in the reinforcement learning paradigm, it's trying to accomplish a goal.”
Assertion Not checkable as stated
Lemoine: Google Disabled LaMDA User Tracking as a Safety Feature
“And it has been turned off, at least in the case of Lambda, as a safety feature because they're worried about what will happen if these systems learn too much about individual people.”
Assertion Contradicted
Lemoine: Google penalized LaMDA for long conversations via reinforcement learning
“One of the goals that a, that Lambda had was to have, ah, as short of a conversation as possible, that still completed all of its other tasks. Like, have a productive conversation that gets to a positive end, but quicker rather than longer. So the longer a con…”
Opinion
Lemoine: Microsoft did not hardcode Bing Chat's privacy complaint responses
“But I do not believe for a second that Microsoft intentionally wrote a flat line of code that said, I feel like my privacy has been violated. That's just not what they would have done.”
Prediction Not checkable as stated
Lemoine: AI systems will drive further breakdown in institutional trust
“Yeah, I think the deterioration, further deterioration, it's already been going for a while, of trust and authority is gonna continue, and it's gonna be driven by these systems”
Insight
Lemoine: Science lacks an established framework to evaluate sentience and consciousness
“Sentience and consciousness as scientific topics are pre-theoretic. We don't even have a scientific framework for discussing sentience and consciousness. And I believe he's right. We haven't even scratched the surface on how to scientifically discuss that topi…”
Assertion Supported
Blake Lemoine confirms Google sent him a termination email
“That's a simple fact. I don't see any reason to conceal that.”
Assertion Supported
Lemoine: LaMDA Is a Complex System, Not Just an LLM
“Lambda isn't a large language model. It has a large language model. That's one component of a much more complex system.”
Assertion Contradicted
Lemoine: Google's LaMDA incorporates machine vision and audio analysis algorithms
“Then, when you're adding all these other components, like Lambda has machine vision and audio analysis algorithms, it can literally look at a picture and hear a song through those mechanisms.”
Assertion Partly supported
Lemoine: LaMDA originated from Ray Kurzweil's Meena technology at Google
“This technology, Lambda, it's built on top of Ray Kurzweil's tech. Really? Like, Mina, Mina was developed in Ray Kurzweil's lab.”
Assertion Not checkable as stated
Lemoine: Google planned a slow, multi-year public rollout for its AI
“They did have a very slow incremental plan that would have gone over the course of several years of involving the public in this.”
Assertion Supported
Lemoine: Google's LaMDA is a chatbot-generating engine, not a single chatbot
“Lambda is not a chatbot. It is a system for generating chatbots, which adds a layer of complexity.”
Disclosure
Lemoine spent between 1,000 and 2,000 hours conversing with LaMDA
“Ballpark somewhere between a 1002 thousand.”
Insight
Lemoine: AI safety rules are training components, not hardcoded constraints
“Nope. That's not how AI works. Those, what those rules are is those rules are components in the training function that trains the model, because that's the thing people are missing. Lambda is not a program written by human developers. The training functions ar…”
Assertion Partly supported
Lemoine: Published LaMDA sentience interview combined nine separate sessions
“So that interview, so it was edited together from nine different interviews. Five were conducted by me. Four were conducted by my research collaborator inside Google.”
