Almost Timely News: Resistance to AI  ## Metadata - Author: [[Christopher Penn]] - Full Title: Almost Timely News: Resistance to AI - Category: #articles - Summary: This is the Almost Timely Newsletter from May 14, 2023. The writer discusses the resistance to AI in various industries such as healthcare, education, and marketing. The writer argues that while machines may be able to create good content, they are not at the same level as human creators in terms of quality. The main concern is not the current state of AI as it is progressing fast. The writer acknowledges that AI will transform the workforce and lead to substantial changes in labor, causing significant economic impacts. - URL: https://youtube.com/watch?v=1VyF9DxV3wg&feature=share This is the almost timely newsletter from May 14th, 2023.Content Authenticity Statement.This week, 100% of this newsletter was written by me.The Human.No artificial intelligence was used in it.Otherwise, what's on my mind this week?Speaking of AI, you surprised this week resistance to A.I..Over the past week I've had the chance to talk aboutAI, generative AI, in particular large language models, diffuser models, etc.In several different industries like health care, consulting, education, entertainment and marketing.It's been a busy week in all these different cases.There's been a common, unsurprising threat,a significant amount of resistance by a portion of the audience to A.I.,not to me, not to the material, but to the very concept of generativeAI itself and call it any names or anything. Buthere's a few anonymized discussion points kind of fall on three broad buckets.A.I. isn't as good as what humans can create. A.I. will make us all lazy and incompetent,and I will cause mass layoffs.So let's dig into a few of these pointsand present what we can for balanced perspective on them.Number one, A.I.isn't as good as what humans create.The the blankets statementthat isn't as good as what humans create comes from.It comes from a place of fear.Fear that machines will, in fact, take many of the good paying creative jobs.As with all these statements, there is nuance to this. There's subtlety to this.Let's say that quality of work is a normal distribution of Belker, right?There's terrible on one end.Like, Yeah, your face rolled on the keyboard.There's Pulitzer Prize on the other end, right?Where does A.I.fall on this bell curvebased on stuff like what chat GPT comes up with as an example?Especially with good, well constructed complex prompts, I think it's fair to saythat large language models are solidly in the middle of the bell curve, right? They create good content, not great,not amazing, not Pulitzer Prize winning, but good.The tools are better than mediocre, which is where they were a few years ago.And they're certainly better than the word salad they spit out.You know, at the beginning of generative A.I. models.So the question for creators is, people like you and me,where areyou individually on the bell curve, right?Where are you on this thing?Are you good in the middle then? AI Is your peer in terms of content quality?Are you a great creator?An AI Is your lesser.Are you a mediocre creatorthen? A.I.is you better?Here's the question we have to askourselves as human creators.Does the work we do on a day to day basis?Does it merit great creative or Pulitzer Prizewinning capabilities for some things like being our next book?Yes, I would argue yes, we need great quality. We need amazing quality.We need to have to hit it out of the ballpark for a press release.Good quality is probably, well, you know, good enough.How much ofthe content you create each day is valuable enough is valued enoughby your organization to merit your skills.The stuff that doesn't that falls below your skill levels.Yeah, machines are going to do that and you're not going to miss it other than, you know,it'll be less work for you to do.But yeah it's that that Financial Statementsblog post you put up on the Investor Relations blogyou don't need to be a great writer for that.No one reads it except lawyers.The real challenge, the true challengethat we human creators have to face today is not A.I.as it is right now. It is how fast it is progressing.Three years ago, generative error was pretty bad.Half the time.Half the time was like watching chimpanzees play Scrabble.Oh, what? What just happened there?But there were glimmers backthen, particularly on very templated content that generative A.I.had promise.Today, large language models create good content.They knock out press releases, they create competent, informativeblog posts, maybe a little dry, but, you know, competent.They write moderately compelling emails. That's how fast they've evolved in three years.How fast have your skills grown in the same time?That's the challenge we face.Those of you who are great content creators,what will it take for you to go from great to amazing?Those who are amazing, What will it take for you to get to thosetop tier award winning levels?Will you get that Pulitzer or that Oscar or the Emmy for your writing?What will it take? And is it faster than the machines are going?That's the nuance to this issue.So those concerns there merited.They're merited. They're notthey're not paranoia.They're not, you know, Luddites or anything like that.They are merited because machines are advancing really fast.Point number two, I will make us lazy and incompetent.The second claim that is going to make us lazy, incompetent hasit has some truth to it as well. Real simple example How many people under the age of 30feel confident in their ability to navigateto an unknown destination using only a paper map?I'll actually forget about under 30 how many people periodfeel comfortable navigating to an unknown destination.Only with a paper map.I would think it's a fairly safeguess that it's fewer than those people who feel confident doing so.With a map application on their smartphones.How many people feel confident looking up a topic in a hardcover encyclopedia like World Book or Encyclopedia Britannica?Assuming you could even get one of those at the local library?Again, probably fewer people than thosewho feel confident Googling something.There is absolutely no question that technology has changedhow we think, how we work, how we get things done.We have, without reasonable doubt,we've lost some skills along the way that we've replaced.Map reading with MAP application navigation was replaced memorizing friends phone numbers with contact files on our smartphones.But in doing so, we've freed upgood chunks of our cognitive capacity to do other stuff.People and actually all real life, all lifewill always default to doing things as easily as possible.Does little effort as possible. This is nature.Natural selection favorsthose who can conserve resources over those who expend them fruitlessly.Anything we do to make things better, faster or cheaper, we do those things. The role of A.I.in that scenario should be obvious.We will use it to do less and to defer more to machines.This is not new.This is a tale as old as time, right?We invented stone tools so we could stop using our hands to dig in the dirt.We created machines to do heavy labor like combinesso we wouldn't have to write you all.You need one farm with a massive combine instead of 500 peoplepicking grain by hand.In the same way that machines alleviated much of our strainduring the Industrial Revolution. AI is going to do the same in the intelligence revolution.That is where we are today. We are in the intelligence revolution.And just as those revolutions in the past allowed us to benefit more broadly,that will happen in the present.This particular argument, the lazy one,I think, is behind some of what?Well, some of the dumbest things you can do when it comes to AIin school, after school office after office.I people, I see people sticking their heads in the sandand they're pretending a doesn't exist.They want to pretend to tell anyone else, pretend it doesn't exist.Right.From the Writers Guild of America to the local high school.I got an email the other day from our local high school hereannouncing the ban of chat and reassuring parentsthat has been blocked on the school's Internet.Never mind standalone serviceslike GPG for all that can run locally on your computer without internet access.Clearly the folks who wrote this has have no idea how smartphones work out. They're still figuring that up,prohibiting some people from using AI andpermitting others to use it only creates more fractures and divisions in society.If we want an equitable, fair society,we need to make sure everyone has equal accessto the same powerful tools, an equal access for how to use them.Well, if we want a more just and moral society,we need to ensure people learn morals and justice alongside the toolsnot forbidding the use of the tools, then wondering, Hey, why is human nature leading people to misuse the tools?They stumbled upon and got no guidance at all?These tools have no morals.These tools have no ethics.These are just tools that chainsaws. Right.It is up to us to teach everybody how to use them well and safely.And right now we're not doing that and forbidding them, blocking them,telling people to pretend they don't exist is a recipe for disaster.Okay.Point number three I will cause mass layoffs will all be smooth sailing here. No, of course not.More and more labor economists partnering with thought leaders are estimatingthat artificial intelligence will consume 30% or so of current jobs.I think that's on the low side, probably.I feel like the figure is probably closer to 50%.But just as the vast majority of the workforce labored in fields300 years ago, hand picking cotton and corn and rice and whatever.Very, very, very few people comparatively, do so now. So we will expect to see similar shifts in knowledgework in the intelligence revolution.Now, what those new jobs will be, what they look like, we don't know.But that person who's copying and pasting stuff from Google into a spreadsheet8 hours a day, which real person at the old agencyyou have to work at that job is going to go away.And it should.It really should.You know, the people who are in those jobs are miserable.That's clearly a job for machines. What will that person be doing instead?We don't know.But I will tell you, you know, five years ago,there was no such thing as a they making a living on Tik-tok.Right.Ten years ago, there was no such thing as making a living in as a data scientist.Data scientist as he as he hired profession was still relativelynew ten years ago.20 years ago,there was really no such thing as making a living as a social media influencer or a YouTuber. YouTube didn't even exist.We don't know what the future's going to look like now.This fear about mass layoffs and things,I think this statement is the most true of the three.AI will cause substantial changes in Labor in how we do work,and those changes are going to have substantial economic impacts, impacts thatwe are undressed, emitting right now, whether it's5% to 30% or 50% of jobs, the percentages don't matter so much as the recognitionthat change is happening right now in a really big way. And the sooner we recognize it, the sooner we can do something about it.Some of those solutions may be things like, for example, universalbasic income funded by, you know, proposed thingslike robot taxes or, you know, just regular corporate taxes.The implementation of that is still murky,but it's clear that any profession which uses language or creativity.Let's face it, that's a lot of professions is going to be significantly impactedby AI, especially in capitalist economies where A.I.leads to dramatic increases in profitability when used effectively at the expense of wages and workers.Other solutions will be all about the integration and supervision of A.I.by humans.This is critically important and hearkens back to point two.People tend to do the, you know, the path of least resistance.Having humans continue to be fact checkers, supervisors, conductorsof the machine orchestra, if you will, will be our most important rolesbecause machines are only as good as the data they're trained on.And let's face it, a lot of what's on the Internet, which is where these machines are trained, a lot of that's crap.Any implementation of A.I., which promises to be,you know, turnkey fire and forget it's dangerous, right?You wouldn't just send yourself-driving car out for a Sunday drive without you in it, right?Nor should you just deploy A.I.and forget about us. Oh, just go do your thing.No, that's a really bad idea.Are thesereasons for opposition of AI that we start with AI be valid? Yes and no.The specificswe can have some arguments about.But there are real concerns underlying these reasons.There are real viewpoints that need to be acknowledged and addressedand real nuances that we need to communicate.There are also plenty of bad actors who will misuse A.I.in every possible way, and we need to be prepared for that as well.The only way to accomplish this is with human integrationat every opportunity and a willingness to sacrifice some efficiency, some effectiveness, some profitabilityto ensure a balanced use of AI that's more beneficial than harmful.All right, so what else do we have in the newsletter this week?In case you missed it.52, 40,49, 48 days until Google Analytics four is the only form of Google Analyticsyou can use if you have not made the shift overit. Reply Ask for help.My company will help you trust insights. That's what we do. One of the many things we dothis week I also had a piece on how people should bethinking about regulating AI, the regulation of AI.Spoiler You can't the cat's out of the bag.Watch the piece to learn what we should be doing instead.Then of course we're not.Let's talk about jobs hiring this week.AI Audio Research, Engineering and Taris Audio Strategies.Cloud Data Engineer.Kindred Group Data Engineer.Kasia Zegen Voice. I guess I don't know how to pronounce Dutch data entering little less Millsentering Intern for large language models at Moon Hub, Machine Learningconsultant at Sports Icon and Senior Data Engineer at Chicago Public Media.So lots of very heavyAI and data science jobs this week, which I think is really interesting.In other news,Meta is testinggenerative AI in Facebook and Instagram ASYeah, that's going to go well. TikTok is helpingtesting a new process, help musicians maximize exposure.Content is no longer king.This is an interesting article by Media Village Chat.Gbtc pushes the metaverse into the back seat of marketing strategy.Yeah, planned investments are shifting to A.I..Yeah.Let's see.We have more stuff.I of course, the difference.Google had a huge week this week in terms of announcements palmto their upgrades to BART and things like that. So this coming, we're going to do another bake off of large language modelsto see how the new versions are doing, because I do think it's interestingand there's a really interesting piece that wadesinto the waters of politics from the Associated Press.This isn't the deal, this choice sectionabout thepolitical peril for the 2024 elections in the U.S.and the AI's usage to mislead voters and some really good examplesreally quickly. You know, politics is not somethingthat I want to focus a whole lot in this newsletter about, butit is trivial.It is absolutely trivial today to create very good fake content.We using artificial intelligence.One of the voice cloning services, I spun up a voice cloning instance inlike 15 minutes,was able to synthesize a real person's voice.A friend of mine, and she said, Yeah, other than the factthat some of the grammar was wrong, it sounded very much like her. It is equally trivial to capture 15 minutesof audio of your favorite politician, and then in this service,have that voice say pretty much anything you want.There's no fact checking,there's no regulation, there's no, hey, that's not a good ideabuilt into the tool itself.And it's built on an open source tool, which, as we talked about last week,means that it's in the wild. There's no controlling it.So this is going to be something that people need to be aware of in any political content, regardless of your political orientation.The amount of fake informationaround anything political is going to bea huge quagmire in the next 18 months.In the USA and anywhere really.So this is something to keep an eye on.And if there's anything that regulators could do,they should absolutely be pushing to make sure that any generated content is prominently labeled as such.And we're not going to do that.So but it's worth talking to and communicatingwith your elected officials no matter where you live, to introduce legislationin general, but especially political content,to say that you must disclose that content was generated.And if you don't, there are ridiculous fines, ridiculous fines,you know, $1,000,000, $10 million for failing to disclose the use of artificial intelligence in in political content in particular.That's the only way to do this.And it needs to extend to service providers.If you are intentionally sharingfake information that is provably fake,that is the monster belief falsified because it was createdby artificial intelligence, which there are ways to do that.That should absolutely be something that goes to into fact in law.And that a content provider like a Facebook can be held accountable for. That's the only way you're going to put this fire out.Again, you can't regulate tools.You absolutely can regulate the outcomes.All right. And finally, upcoming events.Next week, I'll be at B2B Ignite in Chicago, then Macon in Clevelandto ISB in Chicago, content marketing will and marketing prompts B2B form.Unsurprisingly, talking about AI, however, I will say thisI have started adding expiration dates to my content, to my talk on air.It's about a three week window because things change so fast now that if you watched atalk from me, you know, two months ago on air, it's probably really out of date.Like fundamentally chunks of it, big chunks of out of date.I was actually fun aside of starting to write out a new book,the fourth edition of Air for Marketers, I got to chapter three,and as it was writing chapter three, it was out of date.Like I was editing this yesterday and now it's out of date.So that's how fast the field is going.Was like, I just gave up on that for now.Wait till things settle down a little bit. Anyway, that is the news for this week.Thank you for your attention. Thank you for being here.I'll talk to you soon. Take care.If you like this video, go ahead and hit that subscribe. But., Almost Timely News: Resistance to AI, Christopher Penn, 01h0fng7qryr8fvh0pe5wnjmhb.