Sorry I’ve missed some weeks here. A consequence of the holiday and work on a couple of journalistic pieces touching on robots and AI. Please bear with us (actually, me) here as we (actually, I) get some new features ready for this blog — features that should lead to more frequent and varied posts. For now, here are some interesting things about robots and AI for this week.
Insights from Philosophers About Bullshit
It's often said that Large Language Models like ChatGPT, Bard and souped-up Bing are indifferent to reality. They just make plausible sounding texts, without knowing or caring if the words are true. IOW, they're bullshitters.
As it happens there is a growing literature on the philosophy of bullshit. It was launched by Harry Frankfurt in the 1980s, when he pointed out that B.S. (true or false, don't care) is different from lying (I need to know what is true, so I can be sure I'm denying it). I suspect this philosophical work might be useful to ponder in the context of LLM "hallucinations."
For instance, this recent paper, by the philosopher Thorian Harris, might well be applicable to AI users. He argues that we're mistaken to if we focus our anti-bullshit energy only on reprehensible people who intentionally bullshit. Instead, he says, we should worry about bullshit that spreads because we don't bother to resist. (His discussion of Buddhist takes on this problem is a reminder that we needn't look only to European traditions for ways of understanding our AI moment.) h/t The Syllabus
2. Who Needs Women When We Can Get Robots?
Gender equality has come about throughout the world with the entrance of women into the formal jobs-and-wages economy, notes the anthropologist Alice Evans, in a recent post to her excellent Substack.
Women have entered this workforce because their work was needed there. To the extent that robots reduce the need to hire more people, then, it's possible that robots also reduce the social and psychological impetus to favor gender equality. In this post, she makes the case that this may be happening now in South Korea.
3. A Tyrant’s Handy AI Ramadan Detector
Chinese firm admits it won a contract for a surveillance system that correlates dining records and other data with student IDs so that administrators can detect if any Muslim students are fasting for Ramadan. Hat tip: Semafor. link: https://ipvm.com/reports/hikvision-fasting
4. Anguilla Cashes In
It's a good time to be online in Anguilla. Its domain names end in "ai." Looks like it could collect some $30 million this year in domain name registrations from startups and others who want that artificial intelligence vibe in their URL.
5. The Importance of Robot-Makers’ Responses to Regulations
Taxes and regulations on robots are often framed in terms of their effects on demand. That is, what will buyers and users do in response? But regulations will have effects on IT suppliers. Those trying to regulate robotization (for example, with taxes and programs to compensate for lost jobs) should take the (not always obvious, sometimes on the other side of the world) responses of robot makers into account, writes the economist Fabrizio Leone here. Take another, bow, The Syllabus.