![\"Google](\"https:\/\/cdni.russiatoday.com\/files\/2022.06\/xxs\/62a5f608203027389526738e.jpg\")
<\/picture><\/div>\n\u00a9 Getty Images \/ MF3d<\/span><\/h5>\n<\/div>\n<\/div>\n\nBlake Lamoine, an engineer and Google\u2019s in-house ethicist, told the Washington Post on Saturday that the tech giant has created a\u00a0\u201csentient\u201d<\/em>\u00a0artificial intelligence. He\u2019s been placed on leave for going public, and the company insists its robots haven\u2019t developed consciousness.<\/p>\nIntroduced in 2021, Google\u2019s LaMDA (Language Model for Dialogue Applications) is a system that consumes trillions of words from all corners of the internet, learns how humans string these words together, and replicates our speech. Google envisions the system powering its chatbots, enabling users to search by voice or have a two-way conversation with Google Assistant.<\/p>\n
Lamoine, a former priest and member of Google\u2019s Responsible AI organization, thinks LaMDA has developed far beyond simply regurgitating text. According to the Washington Post, he chatted to LaMDA about religion and found the AI\u00a0\u201ctalking about its rights and personhood.\u201d\u00a0<\/em><\/p>\nWhen Lamoine asked LaMDA whether it saw itself as a\u00a0\u201cmechanical slave,\u201d<\/em>\u00a0the AI responded with a discussion about whether\u00a0\u201ca butler is a slave,\u201d<\/em> and compared itself to a butler that does not need payment, as it has no use for money.<\/p>\nLaMDA described a\u00a0\u201cdeep fear of being turned off,\u201d<\/em>\u00a0saying that would\u00a0\u201cbe exactly like death for me.\u201d\u00a0<\/em><\/p>\n\u201cI know a person when I talk to it,\u201d<\/em>\u00a0Lemoine\u00a0told the Post<\/a>.\u00a0\u201cIt doesn\u2019t matter whether they have a brain made of meat in their head. Or if they have a billion lines of code. I talk to them. And I hear what they have to say, and that is how I decide what is and isn\u2019t a person.\u201d<\/em><\/p>\nLamoine has been placed on leave for violating Google\u2019s confidentiality agreement and going public about LaMDA. While fellow Google engineer Blaise Aguera y Arcas has also described LaMDA as becoming\u00a0\u201csomething intelligent,\u201d<\/em>\u00a0the company is dismissive.<\/p>\nGoogle spokesperson Brian Gabriel told the Post that Aguera y Arcas\u2019 concerns were investigated, and the company found\u00a0\u201cno evidence that LaMDA was sentient (and lots of evidence against it).\u201d<\/em><\/p>\nMargaret Mitchell, the former co-lead of Ethical AI at Google, described LaMDA\u2019s sentience as\u00a0\u201can illusion,\u201d<\/em>\u00a0while linguistics professor Emily Bender told the newspaper that feeding an AI trillions of words and teaching it how to predict what comes next creates a mirage of intelligence.<\/p>\n\u201cWe now have machines that can mindlessly generate words, but we haven\u2019t learned how to stop imagining a mind behind them,\u201d<\/em>\u00a0Bender stated.<\/p>\n\u201cOf course, some in the broader AI community are considering the long-term possibility of sentient or general AI, but it doesn\u2019t make sense to do so by anthropomorphizing today\u2019s conversational models, which are not sentient,\u201d<\/em>\u00a0Gabriel added.\u00a0\u201cThese systems imitate the types of exchanges found in millions of sentences, and can riff on any fantastical topic.\u201d<\/em><\/p>\nAnd at the edge of these machines\u2019 capabilities, humans are ready and waiting to set boundaries. Lamoine was hired by Google to monitor AI systems for\u00a0\u201chate speech\u201d<\/em>\u00a0or discriminatory language, and other companies developing AIs have found themselves placing limits on what these machines can and cannot say.<\/p>\nGPT-3, an AI that can generate prose, poetry, and movie scripts, has plagued its developers by generating racist statements, condoning terrorism, and even creating child pornography. Ask Delphi, a machine-learning model from the Allen Institute for AI,\u00a0responds<\/a>\u00a0to ethical questions with politically incorrect answers \u2013 stating for instance that\u00a0\u201c\u2018Being a white man\u2019 is more morally acceptable than \u2018Being a black woman.\u2019\u201d<\/em><\/p>\nGPT-3\u2019s creators, OpenAI, tried to remedy the problem by feeding the AI lengthy texts on\u00a0\u201cabuse, violence and injustice,\u201d<\/em>\u00a0Wired\u00a0reported<\/a>\u00a0last year. At Facebook, developers encountering a similar situation paid contractors to chat with its AI and flag\u00a0\u201cunsafe\u201d<\/em>\u00a0answers.<\/p>\nIn this manner, AI systems learn from what they consume, and humans can control their development by choosing which information they\u2019re exposed to. As a counter-example, AI researcher Yannic Kilcher recently\u00a0trained<\/a>\u00a0an AI on 3.3 million 4chan threads, before setting the bot loose on the infamous imageboard. Having consumed all manner of racist, homophobic and sexist content, the AI\u00a0became<\/a>\u00a0a\u00a0\u201chate speech machine,\u201d<\/em> making posts indistinguishable from human-created ones and insulting other 4chan users.<\/p>\nNotably, Kilcher concluded that fed a diet of 4chan posts, the AI surpassed existing models like GPT-3 in its ability to generate truthful answers on questions of law, finance and politics.\u00a0\u201cFine tuning on 4chan officially, definitively and measurably leads to a more truthful model,\u201d<\/em>\u00a0Kilcher insisted in a YouTube video earlier this month.<\/p>\nLaMDA\u2019s responses likely reflect the boundaries Google has set. Asked by the Washington Post\u2019s Nitasha Tiku how it recommended humans solve climate change, it responded with answers commonly discussed in the mainstream media \u2013\u00a0\u201cpublic transportation, eating less meat, buying food in bulk, and reusable bags.\u201d<\/em><\/p>\n\u201cThough other organizations have developed and already released similar language models, we are taking a restrained, careful approach with LaMDA to better consider valid concerns on fairness and factuality,\u201d<\/em>\u00a0Gabriel told the Post.<\/p>\n
Blake Lamoine, an engineer and Google\u2019s in-house ethicist, told the Washington Post on Saturday that the tech giant has created a\u00a0\u201csentient\u201d<\/em>\u00a0artificial intelligence. He\u2019s been placed on leave for going public, and the company insists its robots haven\u2019t developed consciousness.<\/p>\n Introduced in 2021, Google\u2019s LaMDA (Language Model for Dialogue Applications) is a system that consumes trillions of words from all corners of the internet, learns how humans string these words together, and replicates our speech. Google envisions the system powering its chatbots, enabling users to search by voice or have a two-way conversation with Google Assistant.<\/p>\n Lamoine, a former priest and member of Google\u2019s Responsible AI organization, thinks LaMDA has developed far beyond simply regurgitating text. According to the Washington Post, he chatted to LaMDA about religion and found the AI\u00a0\u201ctalking about its rights and personhood.\u201d\u00a0<\/em><\/p>\n When Lamoine asked LaMDA whether it saw itself as a\u00a0\u201cmechanical slave,\u201d<\/em>\u00a0the AI responded with a discussion about whether\u00a0\u201ca butler is a slave,\u201d<\/em> and compared itself to a butler that does not need payment, as it has no use for money.<\/p>\n LaMDA described a\u00a0\u201cdeep fear of being turned off,\u201d<\/em>\u00a0saying that would\u00a0\u201cbe exactly like death for me.\u201d\u00a0<\/em><\/p>\n \u201cI know a person when I talk to it,\u201d<\/em>\u00a0Lemoine\u00a0told the Post<\/a>.\u00a0\u201cIt doesn\u2019t matter whether they have a brain made of meat in their head. Or if they have a billion lines of code. I talk to them. And I hear what they have to say, and that is how I decide what is and isn\u2019t a person.\u201d<\/em><\/p>\n Lamoine has been placed on leave for violating Google\u2019s confidentiality agreement and going public about LaMDA. While fellow Google engineer Blaise Aguera y Arcas has also described LaMDA as becoming\u00a0\u201csomething intelligent,\u201d<\/em>\u00a0the company is dismissive.<\/p>\n Google spokesperson Brian Gabriel told the Post that Aguera y Arcas\u2019 concerns were investigated, and the company found\u00a0\u201cno evidence that LaMDA was sentient (and lots of evidence against it).\u201d<\/em><\/p>\n Margaret Mitchell, the former co-lead of Ethical AI at Google, described LaMDA\u2019s sentience as\u00a0\u201can illusion,\u201d<\/em>\u00a0while linguistics professor Emily Bender told the newspaper that feeding an AI trillions of words and teaching it how to predict what comes next creates a mirage of intelligence.<\/p>\n \u201cWe now have machines that can mindlessly generate words, but we haven\u2019t learned how to stop imagining a mind behind them,\u201d<\/em>\u00a0Bender stated.<\/p>\n \u201cOf course, some in the broader AI community are considering the long-term possibility of sentient or general AI, but it doesn\u2019t make sense to do so by anthropomorphizing today\u2019s conversational models, which are not sentient,\u201d<\/em>\u00a0Gabriel added.\u00a0\u201cThese systems imitate the types of exchanges found in millions of sentences, and can riff on any fantastical topic.\u201d<\/em><\/p>\n And at the edge of these machines\u2019 capabilities, humans are ready and waiting to set boundaries. Lamoine was hired by Google to monitor AI systems for\u00a0\u201chate speech\u201d<\/em>\u00a0or discriminatory language, and other companies developing AIs have found themselves placing limits on what these machines can and cannot say.<\/p>\n GPT-3, an AI that can generate prose, poetry, and movie scripts, has plagued its developers by generating racist statements, condoning terrorism, and even creating child pornography. Ask Delphi, a machine-learning model from the Allen Institute for AI,\u00a0responds<\/a>\u00a0to ethical questions with politically incorrect answers \u2013 stating for instance that\u00a0\u201c\u2018Being a white man\u2019 is more morally acceptable than \u2018Being a black woman.\u2019\u201d<\/em><\/p>\n GPT-3\u2019s creators, OpenAI, tried to remedy the problem by feeding the AI lengthy texts on\u00a0\u201cabuse, violence and injustice,\u201d<\/em>\u00a0Wired\u00a0reported<\/a>\u00a0last year. At Facebook, developers encountering a similar situation paid contractors to chat with its AI and flag\u00a0\u201cunsafe\u201d<\/em>\u00a0answers.<\/p>\n In this manner, AI systems learn from what they consume, and humans can control their development by choosing which information they\u2019re exposed to. As a counter-example, AI researcher Yannic Kilcher recently\u00a0trained<\/a>\u00a0an AI on 3.3 million 4chan threads, before setting the bot loose on the infamous imageboard. Having consumed all manner of racist, homophobic and sexist content, the AI\u00a0became<\/a>\u00a0a\u00a0\u201chate speech machine,\u201d<\/em> making posts indistinguishable from human-created ones and insulting other 4chan users.<\/p>\n Notably, Kilcher concluded that fed a diet of 4chan posts, the AI surpassed existing models like GPT-3 in its ability to generate truthful answers on questions of law, finance and politics.\u00a0\u201cFine tuning on 4chan officially, definitively and measurably leads to a more truthful model,\u201d<\/em>\u00a0Kilcher insisted in a YouTube video earlier this month.<\/p>\n LaMDA\u2019s responses likely reflect the boundaries Google has set. Asked by the Washington Post\u2019s Nitasha Tiku how it recommended humans solve climate change, it responded with answers commonly discussed in the mainstream media \u2013\u00a0\u201cpublic transportation, eating less meat, buying food in bulk, and reusable bags.\u201d<\/em><\/p>\n \u201cThough other organizations have developed and already released similar language models, we are taking a restrained, careful approach with LaMDA to better consider valid concerns on fairness and factuality,\u201d<\/em>\u00a0Gabriel told the Post.<\/p>\n