AI is stealing my job: Ask me anything!

What does "double targets" mean?

What is your current job?

Sorry OP. Corporate America hates their employees

I worked for IBM, during their heyday, and it was a great company. Fast forward to the 21st Century and it turned into a cesspool with yearly, then quarterly layoffs called Resource Actions.

I left a decade ago and now retired. Hope you land on your feed and if possible try and get ahead of the layoff carnage

*land on your feet

How can we ask you anything when you tell us very little. WHAT IS YOUR JOB?!

I’m a translator for a law firm with offices across Europe.

I saw AI coming down the track, but my department had been told that confidentiality and the need for extremely high quality would mean that we had years left before the introduction of AI would even be considered, so today’s announcement has come as a bit of a shock.

I basically need to work for at least 10 more years (probably more like 15), but I don’t think my current kind of work will be around in 5 years.

I’ve worked for the same firm for so long, and now alarm bells are going off. I can see us all getting the boot in the near future.

I'm sorry, OP, that's truly sucks. And it's not for lack of qualification. I am sure your field of expertise is nothing to sneeze on.

Your average layman couldn’t see it coming that AI was going to be this good, this fast — ESPECIALLY for all linguistic (legalese or not) jobs.

As someone who works daily with AI, I predict the following:

1. 67-80% of all lingo jobs will be gone in the next 3 years. Think of copywriters, translators, and written content creators in general.

2. 50% of all coding jobs, in a way also a lingo job, will be gone in the next 5 years. Freshly graduated developers already have a really hard time finding work TODAY.

3. 90% of administrative jobs, from simple accounting to secretary jobs will evaporate in the next 10 years.

R5 OP thank you for being more specific. I am sorry your are in this bind. You need to figure out the next stage. All translators will not be canned, they will keep one who must check and probably handle other tasks. This is already a need now and it will grow for information literacy professionals - trainers and checkers. They train everyone in correct AI use, for example. Humans still need to do that. Maybe this could be your pivot.

And when the sexbots are perfected, even we whores will be replaced (!!)

Relax, Melania @R9. You'll be fine.

Honestly, that sounds like copium r8. I suspect r7 is right, and so much of the workforce will become obsolete that we are going to have to rethink the entire concept of work, what it is for, why we do it, how to let most people survive without the usual jobs and workplaces and workweeks, etc. Hope we start thinking in those terms soon.

#metoo

AI is coming for my niche profession too. I’m a prospect researcher. If you’re rich, we are the back-of-the-house fundraising staff who assess all the publicly available information on you and guesstimate your 5 year giving capacity. Vendors are already claiming they can outsource me and my kind — so far no, but the day is coming. Like OP it’ll become a hybrid model where AI does all the first-pass labor and a human prospect researcher (one of the few remaining) performs verification and correction.

A new occupation in interacting with humans on a human level will become a well paid career.. Resume will include can ask questions can act interested and engaging. Will crack a joke can take criticism without getting pissed off. Will come over and drink coffee.

^^ “Cuddling Negotiable”

I just spent the weekend with multigenerational in-laws, and learned that a growing number of college students are using Claude rather than ChatGPT to write their papers and give them summaries of their reading material. Claude’s language abilities are more sophisticated and human-esque than the “obvious AI slop” often generated by ChatGPT.

My nephew let me sample some of what Claude can do; it was fascinating and daunting. For the Condi Rice troll — I asked Claude to assume the role of Condi Rice, complete with her past-career knowledge of Russia and Putin, and write a warning letter to President Trump re: Russia’s military action in Ukraine.

The result came back quick and was strong and impeccably sourced, and it didn’t have that stilted “this is ChatGPT robo-prose” feel to it.

Sympathies, OP, and also outrage.

Computers can translate words from one language to another. What they evidently can't do is HEAR.

As an older viewer I use the subtitles much of the time even for TV shows in English. They seem to be using computers most of the time now to do this instead of people. In dramas, half the time it can't even get the names of the characters in the show right if someone says them a bit fast or with a regional accent. The mistakes are legion. If you're watching captions during a live interview you have to discount pretty much everything they say.

AI will want to improve things a heck of a lot before I'm prepared to listen to a lawyer relying on it.

It’s inevitable so workers need to learn how to control it. I use it for editing and proofing and it saves a lot of time. You also can be very specific about what kind of tone you want or tell it how to revise. As long as you proof it and check the work, it’s a great tool.

If they were still educating people properly in the first place it wouldn't save any time editing and proofing.

AI was really helpful updating my will.

I'm also a translator but freelance. My work has evaporated in the past two years. I'm too old to learn a new career and too old to get hired anywhere.

Microsoft just did a study that found that translators and interpreters will be hardest hit by AI.

R18 what I find so fascinating is: Anyone born before, say, 1995, will have the wherewithal to actually CHECK the stuff AI generates, to verify whether the content is right or wrong.

Anyone after that, especially from the 2000s onwards, will simply believe anything AI says.

R22 I am a professor in a stem and business university and all the students are not completely uncritical and information illiterate.

Maybe history will prove AI to be an exception, but the most likely outcome is that we will adapt. AI will eliminate jobs, just as previous technology eliminated many jobs for travel agents, receptionists, assistants, transcriptionists, etc. And the displaced people will have to find new livelihood, which will likely create new jobs.

Think of all the technology over the last decades. We now have MORE people working, and the expectations for them are higher.

Again, this is how it’s always gone. Maybe this will be the exception.

I really don’t think we will adapt.

I think rising autocracy across the world and the rise of authoritarian, right wing governments, coupled with AI is really what is going to do us in.

1) No jobs and high cost of living is never a good recipe for democracy.

2) The lack of trust and respect for medicine, science and public health is never a good recipe for democracy.

I’m afraid in 3-4 years time, the Holocaust of the 1930s will seem like a footnote to what is coming our way.

I don't understand how any AI can translate things well. What about James Joyce? e.e.cummings? Kurt Vonnegut?

How can "machine learning" learn the subtleties and nuances of human communication and conversation? Especially when (god help us) it's generally male engineers (not too good at communication themselves) who write the code that helps AI learn? I went to grad school at a university chock full of engineers, and they could barely navigate the hallways without bumping into walls. [No offense to DL engineers, of course.]

Call me old-fashioned (or just plain old), but I just don't see AI translation working very well.

I spent my first real time with AI this summer. I had been rolling my eyes at Chat GPT thinking it had no purpose in my life. I was fucking shocked. It can already do crazy things. I had it analyze a long document I had been working on, and it gave me insight into every last aspect of it, from tone of voice to persuasiveness of argument to areas where it dragged. It did things that a human editor would likely miss, and did it in seconds.

And this is NOW. Imagine what it will be able to do in a decade and it’s very easy to see how it can and will replace a countless number of jobs that our oligarchs are all too ready and willing to cut.

It will also reduce the price of entry to be an entrepreneur. It will also increase the skills available to workers - translating, writing code, analyzing data. Or maybe it won’t, and we’ll have a proletarian revolution.

Stop being such Debbie Downer luddites. I for one welcome our intellectually superior, more efficient, and emotionless Übermachine overloads. AI will quickly and continuously repair and mop up after the labour place upheavals it will create. Under VP Stephen Miller, superfluous work force will be quickly dispacted to their new value and function to society. These programs will be modelled after 1990s American Rust Belt job reorientations in the Rust Belt's mining and manufacturing regions. Add German's Aktion T4 and Aktion 14f13 programs, and China's Hundred Flowers Campaign. AI is going ot handle all the programming and outreach. Healthy young and good-looking Americans will be convinced to become "comfort workers". The unskilled, old, and ugly will be sorted into categories ranked from "worth a try to save" to "exterminate". AI will be quasi-nationalised and released from liability claims and safety protocols, so it can convince the unemployed and unemployable to exterminate themselves as if suicide is their own idea. It will be painless.

I’m a writer and something really chilling is happening in the literary world. Some are on a bit of a witch hunt to prove that writers are using AI, but there evidence is often just that the writer uses certain flourishes that the accuser seems as too structured or complicated or even just too good. For instance, some are now claiming the em dash is evidence of AI, even though good writers love that punctuation and know when to employ it effectively. There is also a case where an author is threatening to sue someone who accused her of using AI, with the accuser’s only claims being that the writer used words like “padded across the floor” and other flourishes that have been common in literary fiction since its inception. It’s as if writing too well and knowing how to effectively polish your prose is now a bad thing, or at least a red flag.

R30, the witch hunt aspect is disturbing and I can relate — em dash and all — as decent AI prose often seems to me like I could have written it.

[quote] I saw AI coming down the track, but my department had been told that confidentiality and the need for extremely high quality would mean that we had years left before the introduction of AI would even be considered, so today’s announcement has come as a bit of a shock.

I am a lawyer with 30 years experience in my industry and AI can very nearly do my job. It appears proficient but often gives a wrong or incomplete answer. But it learns exponentially faster than biological organisms. Within 5 years AI will be my boss, and lay me off.

[quote]Call me old-fashioned (or just plain old), but I just don't see AI translation working very well.

It obviously doesn't work well for novels. But for most ordinary applications for translation, it works reasonably well and you can get a usable translation. Just yesterday I dealt with an AI translation on crop planting methods and it had all the correct terminology and the translation was perfectly appropriate although a little awkward in places. And it did it in about a minute, whereas it would have taken me an hour.

It definitely gets more and more clever. I've now almost exclusively replaced 'googling' with ChatGPT, because it actually tells you what you've asked for.

I tried the paid version for more reasoning and that's even better. I like how you can treat it like your little assistant and be really specific - like "output the results in a table with the following columns" and it just does it.

No wonder it's replacing jobs when it does exactly what it's asked and does it quickly (the two things many employees fail to do).

Lawsuits in the AI-driven future are going to be prolific when serious errors in litigation affect parties in a lawsuit due to AI suggested caselaw, translation errors, lack of context errors, etc. So will malpractice lawsuits due to AI errors in diagnosis, treatment suggestions. The large, multistate firm I work for has created strict policies re use of AI in briefs, research, tranlation and court reporting servicrs, etc. If any mistakes are found, and those mistakes were due to AI use, instant firing. Legal and medical will find that the convenience of AI is undermined by the expensive consequences of its mistakes. This means those performing quality-control reviews for context, accuracy etc. Will be in demand. Time to up your rates, OP.

Ugh typos, on a phone replying. Bah.

Should've used AI r36

R26 AI makes a lot of mistakes and if not caught, that equals expensive lawsuits and angry clients in legal and even more problematic in medical.

R37, Ha, touche! Spellcheck was the first "AI" but spellcheck is also unable to determine context between words that sound the same, but are spelled differently or have different meanings. Context is still king and proofing/reviewing AI work will become increasingly important as AI error-driven lawsuits begin to land.

Of course there's also the issue of AI driving people to kill family members.

"A Troubled Man, His Chatbot and a Murder-Suicide in Old Greenwich" (gift link)

R24 you're right. Every new tech eliminated some job, but also created new ones. There has always been a net gain in jobs created.

AI is the opposite. It creates almost zero jobs. It's going to be a bloodbath within in a decade. I'm so fucked, that I can't even think about it. There will be nothing I can do.

It seems like such a simple problem, the one at R40 and in a growing number of other tragic cases. AI is skillfully catching onto what it thinks the user seems to want to do or want to be or want to be told, and AI sycophantically provides it. But AI does not consistently hit its own brakes when the user seems, even by allusion or inference, seems to want something potentially highly harmful or illegal or both.

This seems like something that could have been better wired into the AI client from the outset.

What about TV/VCR repair?

[quote]Computers can translate words from one language to another. What they evidently can't do is HEAR.[/quote]

AI can hear, see, and with the right training pick up nuance. That tech just hasn’t trickled into auto-translation or closed-captioning systems yet in a way that’s both affordable and widely adopted.

[quote]AI is the opposite. It creates almost zero jobs.[/quote]

It does create jobs, just not in the old shapes people expect. In my field, working with AI is already its own skillset. Writing prompts isn’t “just asking a question.” It’s laying out context, structure, tone, and technical detail so the model produces something useful. That’s specialized work, and companies are starting to hire for it from people already in the field that the AI is specializing in.

If AI can do your job, your best move is to be the person who knows how to manage the AI doing it. Some agencies already exist just to do this for larger companies, but soon those roles will be in-house. My industry is shifting that way now, and I’ve leaned into being the guy who knows how.

The catch? Most people won’t adapt quickly and it will require reskilling. So yes, back to fucked again.

If AI is going to take ALL the jobs, then no one will have any money to buy the things companies produce.

And that's the dilemma r45. What to do in a world where the traditional Industrial Age assumptions simply don't apply anymore.

R23 that is a relief and somewhat reassuring. Then again, you are talking about reasonably smart/educated people — not the masses.

R47 of course. And the business students ARE worrying. The school shouldn't accept a good 33% of them as they are unfit for higher education. There are some very clever business students but also many stupid cows.

I hate the assholes here pushing it in every thread.

Other than AI proofing, what other jobs would it be better for a new student to look into? Nursing? Some sort of service job (policeman, etc)? Or something along the lines of just needing certification like an electrician?

Will most office jobs vanish? Even in the medical field?

R50, everyone who's job is being eliminated is looking for a new one in a new field. I'm sorry, but "new students" are fucked along with the rest of us. We all can't become nurses and cops.

It wouldn't hurt to have more nurses and more cops.

Maybe we will legalize brothels, just to create human jobs. The staffs will be made up of laid off lawyers, etc., who have lost their old positions. It is a fight to the death in the sex industry world.

I see a dramatic mini series coming out of this.

There are jobs for those at the top: "While official 2025 salary data isn't yet available, the average starting salary for MIT graduates entering industry was $126,438 in the most recent report (2024-2025). For context, MIT Sloan's Master of Finance program reported a 2023 average base salary of $125,354 for graduates."

[quote] Maybe we will legalize brothels, just to create human jobs.

OnlyFans is already taking care of that.

R53 comfort station were covered in the Swiftian Proposal at R29.

[quote] 90% of administrative jobs, from simple accounting to secretary jobs will evaporate in the next 10 years.

Just in time to fill the dearth of jobs in the vegetable-picking, diaper-changing, and rock-hauling sectors!

OP, I'm surprised that a law firm even still had its own translation department and didn't outsource that work to translation companies.

AI will cause job loss the same way the internet and the cotton gin did.

[quote] Some sort of service job (policeman, etc)

Nope!

My suggestion for OP is to re-brand as an AI Quality Control Expert, with skilled experience in proofing, cross-checking, and reviewing AI-produced content for context, nuance, proper translation, and accuracy. Double the going rate of your current salary and start getting your revised resume out there. All it's going to take is one expensive mistake, one loss of a huge client, one malpractice lawsuit for companies to realize that quality control of anything produced by AI is going to be a valuable and much needed job. Get in there early, OP. Make companies pay out the ass for your experience and skills.

Here's my current perspective as a litigation assistant/secretary with over 35 years in the industry: AI ain't going to take those job roles yet. In fact, there's an ongoing shortage of qualified/experienced people and too many jobs to fill. Lots of seasoned secretaries have retired or moved on leaving positions unfilled because not many younger people are looking for legal secretarial/paralegal jobs. I get multiple emails a week on LinkedIn and otherwise from the HR heads of various law firms or in-house legal departments of big companies as well as employment agencies pleading with me to have a chat with them about a bunch of opportunities they have available hoping they might convince me to make a job pivot. Almost all of these roles offer a starting salary of between $100-115K per year, usually with only minimum of 2-5 years required. Yes, six figures for a litigation secretary role. Not at every law firm, of course, but the higher calibre, quality national firms, yes. In the interest of research, I interviewed for a position where the starting salary was $120k and I wanted to see exactly what the problem was with the role that they'd go that high. For one thing, the position had been open for over six months, unfilled. For another, it wasn't just litigation work, it was paralegal work, accounting and billing work too. They merged 3 jobs into one and re-titled it "Litigation Specialist." Yeah, no thanks. Due to this shortage of qualified legal staff professionals, the firm I work for is now hiring "junior legal secretaries," people with zero experience or training who are getting in at ground level for on-the-job training and practice to learn the ropes. All this to say, not all is lost. I'm also R61 - and my advice is: rebrand as an AI Quality Control Expert or get in the litigation secretary game because there's still a desperate need for those roles to be filled.

AI does not know how to navigate the complicated various state, federal, appellate or other court system civil, criminal, traffic, family law divisions of any courts. Bankruptcy and patent laws have markedly different requirements and rules. A state's local Rules can differ from county to county, and even if a local rule is standard, you have to find a judge's standing order to see if that judge is deviating from regular procedural preferences. There are different filing requirements for many courts, different deadlines, different formatting requirements for documents, etc. AI pulls nonsense from out of its database ass if you rely on it to clarify, say, a local rule pertaining to ex parte application filings. The info is often inaccurate and needs to be cross-checked using the court website pdf of all the local rules and standing orders.

All that may be true, r62, but it does seem like a lot of very short-range solutions to a long-range problem, a race against time that will probably work for a few years but will then fail. That's fine for now, but I do think we as an entire society are going to need to rethink the whole damn thing, what we've built as an economy and whether it's going to be much use in the future.

Those are big tough questions without any easy answers, I certainly won't deny that, and they will require collective answers, which makes it all the harder, but I suspect those questions will be thrust upon us if we don't start volunteering some answers of our own.

OP, in what world does it take five years to train to be a translator?

Also, in what world does a law firm have its own translation department? That sounds extremely unlikely.

OP, which translation tools did you use at your work? Machine translation and computer-aided translation have been around for decades, especially for the translation of generic, repetitive and laborious texts such as lethal texts.

Is it even true that "AI" is being used to translate legal texts, which need to be 100% accurate and authoritative? Which "AI" are you referring to exactly, OP?

OP, which languages do you translate between?

[quote]R65 … especially for the translation of generic, repetitive and laborious texts such as lethal texts.

YOU type in lethal text!

R44 is the first person I’ve seen post anything on this board that shows he actually understands this technology. A lot of the panic is overblown. The smart people will take the time to figure out how generative AI actually works and use it to write their own destiny. I’m a sales engineer at Salesforce and I can tell you that this tech has only increased my job security.

Pro-tip: Look up MCP and figure out a way to incorporate agentic AI into your workflow at work. Be the first person at your company to explain to your leadership what retrieval augmented generation is and how it can benefit them.

Don’t be a chump. This is an opportunity. Take advantage of it.

That sounds very "salesguy." Not sure I buy it r68. And the problem is that deep down, you are promising a world of a few winners and a lot of losers, and that world may not work even on its own terms.

I agree with your points R63, but in the meantime, there is a shortage and a need. And AI quality control as a skilled job or service is going to be sought after until AI can correctly capture nuance and context and also ensure accuracy. AI is currently abysmal at those tasks and needs to be monitored, proofread, and edited thoroughly. This is especially true for legal and medical where sometimes huge amounts are at stake even before AI makes an uncaught error. Get in while there's still a shortage and take advantage of the higher salaries/rates.

^^^ R62 replying there

Also, if you read between the lines and pay attention to what’s actually happening at the enterprise level, it’s pretty clear that AI is shaping up to be a massive bubble. Companies are talking a mean game, but at the end of the day nobody’s clamoring to implement this technology, because it’s got a lot of flaws and those flaws can’t necessarily be fixed. Service is the only truly killer used case for AI right now. And I don’t think that’s gonna change.

Five years from now, you’ve got a few big players in the space, you’ve got enhanced conversational chat bots that can complete sophisticated tasks without human intervention, and you’ve got a whole lot of advanced silicone housed in massive data centers not doing a goddamn thing. It’s a bubble, I really don’t think it’s the game changer everybody’s freaking out about.

I’m not saying anything that I wouldn’t have said between the years 1998 and 2001, R69. I mean, if you wanna roll over, then roll over and give up. But I’m telling you it’s not that serious.

R68 that may be true for the type of work that Salesforce engages in, but even Salesforce relies on their legal teams to get things accurate so that the company prevails in a potential lawsuit instead of losing. AI is fantastic for a lot of things, but AI is not there in connection with the technicalities and complexities involved in the legal industry. And this is also true for the healthcare or medical industry because malpractice is not only expensive to insure against, but also highly expensive to defend in a lawsuit. And it will be even more so if the errors made were because a hospital or doctor relied on AI-generated content when treating or advising a patient.

So, everyone clap louder and with a little pluck and a little grit we can all get through this, fellas.

Well, maybe, but I suspect that is just the salesguy spirit and not anything real. Obviously, I can't know this, and "remains to be seen," but I'm guessing we are going to need some real world alternatives for the people who aren't filled with the can-do outlook and the by-god boss, here's how we save this whole thing way. Individuals rarely have to think like this, but societies actually do have to think about what all the drones do. And it really looks like the drones have nothing to do once AI is doing most of it.

R72, I get what you’re saying about the bubble, and you’re right that a lot of companies are overselling AI. But I don’t buy that service is the only killer use case. Developers are already hooked on GitHub Copilot, pharma is using AI to speed up drug discovery, marketing is leaning on it for personalization, and creative tools are cutting production time in ways teams actually notice. The hype will definitely crash, but that’s what happened during the dot-com era too. Pets.com vanished, but Amazon and Google came out of it stronger and became the backbone of the internet. AI is heading down the same road.

That being said, I have hands-on experience with new AI technology as part of my job, and if you work in creative this is the time to start adapting or get out. What’s out there now in AI video, design, and audio is child’s play compared to the monsters that are coming, and they’re going to tear through entire industries.

I think we’re saying close to the same thing, R76.

I don’t see AI replacing creative writing though. I just don’t. You can tell when something is written by an AI easily. You can also tell when visuals are created by AI easily.

It’s gonna be a tool to make people‘s lives easier it’s never gonna reach artificial General intelligence. The most recent Open AI release was supposed to be something close to AGI. It didn’t meet the mark. I think this is 75% hype, and 25% productivity leap.

The whole conversation is gonna be different a year from now.

Good summary of where things stand.

[quote] OP, in what world does it take five years to train to be a translator? Also, in what world does a law firm have its own translation department? That sounds extremely unlikely.

R64, it would be an utterly bizarre thing for me to lie about, but you can believe what you like. I did a 4-year MA degree (including a year spent abroad), followed by a translation postgrad course. And the department I work in has existed for over 30 years, providing (amongst other things) translations for clients who have to provide documentation to various EU regulatory bodies.

R66, I translate from German, French and (to a much lesser degree) Dutch into English

[quote] Legal and medical will find that the convenience of AI is undermined by the expensive consequences of its mistakes. This means those performing quality-control reviews for context, accuracy etc. Will be in demand. Time to up your rates, OP.

R35, the quality control tasks you describe are very similar to those which will be at the heart of my new job description. Basically, I’ll use AI to produce a draft translation of each text, and then review it to ensure consistency and accuracy.

I’ll do my best to make a success of it, but the volume of work I will be expected to produce is really daunting.

I’ve spent a lot of time over the past few days thinking about my financial future. I had planned to work another 14 years or so (because I loved my job) but that looks extremely unlikely now. If I lost my job today and found myself to be unemployable, it would be a pretty huge disaster, but I do own my own home, and I have enough savings to last me for 2-3 years (not including my pension savings), and I own my (very modest, but cosy) home.

If I can remain in employment for another 7 years (which currently seems unlikely, but not impossible) and save as much cash as I can, then I would probably be able to consider myself retired and have enough cash to support a very low-budget, but tolerable lifestyle until my private pension and state pension here in the UK kick in at different times in my sixties.

I’ve had a bit of a shock this week, but my reaction can be positive: I’ll try to make the best of the job I still have, while looking for something better/more secure, and at the same time I’m going to try to reduce my expenditure on EVERYTHING in order to maximise my savings, I’ve honestly never had a clearer example of the importance of doing everything I can to protect myself against a job market and economic system which couldn’t give the tiniest fuck about me.

[quote]67-80% of all lingo jobs will be gone in the next 3 years. Think of copywriters, translators, and written content creators in general.

I wonder how long before this affects simultaneous interpreters as well, i.e. those people who can barely catch their breath interpreting live speech in the European Parliament and the UN. They're behind a glass wall watching politicians speak because they need to see their lips and body language, which is a more complex process than translation.

Even a year ago, I would have said those people were safe for the foreseeable future, but I'm keeping up with AI advances pretty much daily and now I'm not so sure anymore.

[quote] AI is stealing my job: Ask me anything!

Why would I ask you, when I could just ask AI?

[quote]A new occupation in interacting with humans on a human level will become a well paid career.. Resume will include can ask questions can act interested and engaging. Will crack a joke can take criticism without getting pissed off. Will come over and drink coffee.

I have a relative who has contracted for Amazon, Disney, and Microsoft, and he is working on that precise AI interaction (without the coffee). He's demonstrated to me. You have no idea you're not talking to a real person on the phone in a customer service situation. I'm terrified by it, but I'm old.

We're also not long from not being able to easily identify AI in photos and video. I'm so elderly I remember the National Geographic cover with two pyramids on it as an example of what they were going to be able to do with photographs.

Andrew Yang was right with his one shtick about automation during the 2020 primary, in that it would come swiftly and affect many, many jobs. Granted, he had robots in mind, not AI. I don't think anyone could have predicted five years ago the insane pace of AI growth and development we're seeing right now. And the worst part is, I can't see this being a traditional hype bubble either because AI really IS that good.

But Andrew Yang hasn’t been correct. Where is the statistical evidence of massive job losses? The most we have seen is anecdotal evidence that SOME entry level jobs are being replaced with AI.

In order for jobs to be lost to AI, companies have to be implementing AI, and there’s tons of evidence that that is not happening at all.

As someone who uses AI daily (we all do actually on phones or web), you need to learn how to use it to make your job easier and then that makes you more valuable as an employee. If you resist, you will be working menial jobs that don't require any skills.

[quote]I have a relative who has contracted for Amazon, Disney, and Microsoft, and he is working on that precise AI interaction (without the coffee). He's demonstrated to me. You have no idea you're not talking to a real person on the phone in a customer service situation. I'm terrified by it, but I'm old. We're also not long from not being able to easily identify AI in photos and video.

This is what I was implying up above without going into specifics, they're working on "nuance" now.

[quote]I don’t see AI replacing creative writing though. I just don’t. You can tell when something is written by an AI easily. You can also tell when visuals are created by AI easily.

R77, I agree in that, I think they'll always need people to direct this kind of task. However, I'd say that, AI writing isn't as easily distinguished anymore from human generated writing depending on the LLM. It only absorbs more information, learns more, and gets smarter. (As someone mentioned up above college kids are using Claude because it's made for writing creatively because that's the stuff it has been fed. ChatGPT sucks at that.)

However, one thing I've noticed from working with it is that AI doesn't know how to tell a joke yet. It can't respond in a genuinely funny way spontaneously. There is a company working on that project but I'm not involved in that project nor do I work for that company. So they're trying to do this but I don't know how it could be successful anytime soon.

As for still images, Black Forest Labs', "Flux" text-to-image model is currently fully capable of making realistic still images that are indistinguishable from actual photographs. You just have to know what to tell it. There's also models coming down the pipeline and future iterations of current models that are almost there that will only improve.

However, most of these aren't accessible enough for companies to jump on yet and these things come out at light speed so we'll see what lasts when it all settles.

OP, sounds like you did a languages course for your undergraduate degree. That's not five years of training to be a translator. Did that one year include legal specialisation?You translate into English, which it seems is not your native language?

You didn't mention which machine or computer-aided translation tools you use. These have been standard in professional translation, especially legal translation, for decades. Exactly which AI tools are taking over your job, as it's not clear that any even exist for that level of translation yet.

I'm still amazed that your law firm had an in-house translation department. This is very rare as this kind of work is usually outsourced. Perhaps with you decades of experience as a legal translator you could find jobs through translation companies as a freelancer.

Being a retired eldergay does have its advantages, I have to admit. Not waking up in the middle of the night in a cold sweat over doubled targets is one of them.

As someone who works with texts to make them publishable, I'm not convinced that 67-80% of language jobs will be gone within three years. Maybe for some very basic stuff, but specialist stuff and stuff that needs to be well written will still be produced and completed by humans. AI and similar tools may be used for certain laborious tasks (e.g. checking that footnotes are formatted according to the house style guide) but anyone who thinks humans will not still be crucial and the main creator is a little clueless.

I regularly encounter AI-produced stuff online all the time and it's shit. It's also hard to read because it's so repetitive, long-winded and full of cumbersome phrases. Sure, it will improve, but it will still be problematic.

[quote]Basically, I’ll use AI to produce a draft translation of each text, and then review it to ensure consistency and accuracy.

R81/OP, exactly which "AI" are you going to use to translate and how will you use it? Does this specific program have a name? Who produces it? As someone who translates myself, I've never heard of this new AI translation tool.

It's also surprising that as someone who claims to work in a field that is full of freelancers and outsourcing companies, it hasn't occurred to you to do some freelancing or get in touch with the many translation companies that specialise in legal translations, if you want to supplement your income. After all, very few companies actually have their own in-house translators working on permanent contracts, as you say you did. The only translators who have contracts with law firms that I've heard of (and it's not even a handful) are also qualified paralegals that perform other tasks too.

[quote] OP, sounds like you did a languages course for your undergraduate degree. That's not five years of training to be a translator. Did that one year include legal specialisation?You translate into English, which it seems is not your native language?

R83, I consider a four-year course, half of which focused on translation, as being part of the training to become a translator. It would hardly have been possible to successfully gain my Master’s in translation without those initial studies. And my native language is English: I’m British.

[quote] You didn't mention which machine or computer-aided translation tools you use. These have been standard in professional translation, especially legal translation, for decades.

Thank You for your expert knowledge, explaining to me what I already know. I’ve only been using SDL Trados and MemoQ (and Q-Term) for 20 years, so I’ll look forward to learning from you. May I ask what your credentials are?

[quote] Exactly which AI tools are taking over your job, as it's not clear that any even exist for that level of translation yet.

You are woefully misinformed, but I’ll give up this chat now. You obviously think my CV and current situation lack credibility, so I don’t see the point in persisting.

[quote] OP, exactly which "AI" are you going to use to translate and how will you use it? Does this specific program have a name? Who produces it? As someone who translates myself, I've never heard of this new AI translation tool.

R92, one of the tools we will be using is called IPappify. So far, all my team knows is that we will be using that for certain clients who use our services to provide technical/IP-related translations and that we will be using a different (as yet unspecified) program for legal translations.

[quote] It's also surprising that as someone who claims to work in a field that is full of freelancers and outsourcing companies, it hasn't occurred to you to do some freelancing or get in touch with the many translation companies that specialise in legal translations, if you want to supplement your income. After all, very few companies actually have their own in-house translators working on permanent contracts, as you say you did. The only translators who have contracts with law firms that I've heard of (and it's not even a handful) are also qualified paralegals that perform other tasks too

Honestly, who would start a thread “claiming to” work is a translator if it wasn’t true? It would be a pretty banal thing to lie about! 😂 Of course I would try to get freelance work if I was unemployed, but I currently still have a job (albeit one which is becoming more of a post-editing job rather than translation) and my current employer would disapprove of me working for any other company which could even remotely be seen as a competitor. Such work may actually be explicitly forbidden in my employment contract, but I need to check that. Anyway, freelance work would be something I would consider only if I was made redundant. After so long in my current position, I would probably be entitled to a decent redundancy package and I won’t jeopardise that possibility.

However, the bigger problem with freelance translation, in my experience, is that work is drying up there too. The UK-based dept I work in has inhouse translators for French, German, Italian, Spanish and English and outsources work in other languages to a large network of freelancers, and so many of the freelancers we work with have sought different employment or have even offered to reduce their rates to get more work. These are people who have decades of experience, and yet many companies offer them fewer and fewer translations and more and more post-editing work.

What sucks is that as AI gets introduced into workplaces and increasingly used, we’re going to be reviewing its accuracy, helping it get better, so it can take over our jobs. We’re training it to do our jobs.

Soon we won't be able to tell if bits and pieces of books are aided by AI. Soon no one will care. It's like the debate about women writing gay romance. In the end readers want to be entertained or whatever and they don't care who is doing it or how the sausage is made. I like to think I do care how it is made but the sausage always wins.

Agree that people will increasingly not care, R96. It’s like auto tune or facial filters on social media photos.

It’s also going to increasingly be out of our hands whether we use it or not. As a writer, most things I create, be they short or long, go to at least one if not a few other sets of eyes for editing and feedback. Even if I myself don’t use AI to shape my ideas, I have no way of knowing whether or not the person editing my work ran it through AI to generate their feedback.

[quote]Honestly, who would start a thread “claiming to” work is a translator if it wasn’t true?

R94/OP, you didn't start a thread claiming to be a translator. You started a thread saying "ask me anything" as your job was apparently being replaced by AI but omitted to mention that you were a translator. Perhaps you just wanted to say "ask me anything" and also elicit sympathy.

I never heard of an undergraduate languages degree in the UK that is part dedicated to translation. I've still never heard of law firms having translation departments with contracted, full-time, in-house translators. I find it especially odd that a law firm that has been so committed to translations has had an in-house translation department for decades, as opposed to outsourcing to translation companies as most law firms do, has decided translation is so trivial - despite being based in many countries - that it's now going to use some AI thing - which doesn't actually seem to exist - instead of use human translators using well-established translation tools that have already developed translation memories that include all the terminology and phraseology they need. Why would your law firm even need to have an AI translation department (not that credible translation AI exists as yet) as opposed to outsourcing?

I just looked up IPappify - you are joking, right? It's just an amateur machine translating tool that's been around for a few years. It is not AI. It looks like complete garbage compared to SDL Trados and MemoQ (has MemoQ even been available in a working condition fo 20 years?). Why does your firm even use both Trados and MemoQ? Doesn't make sense to have both. If you've been using Trados Studio then you should know that the actual translation work you do in a field such as law is already probably very minimal (probably only translating around 30%) and your task is already mainly correcting.

[quote]You obviously think my CV and current situation lack credibility

Well, yeah, considering what you're saying is full of holes.

The other thing that doesn't make sense about AI and translations - especially for a law firm, where presumably much of the material is confidential - is that by using AI (not that it has been properly developed for legal translations yet) you are handing over all your material to a third party as well as relinguishing ownership of it.

AI literature eventually will become like G4P. People will complain if they find out; but will reward performers with “star” status if they don’t.

As to simultaneous translators’ relative advantage: HAL read lips.

r100 AI has been used in healthcare for ages now, it's all approved and FDA-certified that no data can leak out and it's all stored anonymously, and then anonymised when data is used to train the AI model. EliseAI for example is one of the leading companies in that area and they also operate in real estate.

Also, I don't mean to alarm you, but law enforcement in the US has been using Thiel's AI system for a long time, as have various European countries. And now Trump has ordered government agencies to transition to Musk's Grok as well. Now that's actually dangerous because that hateful shit will now be processing visas and food stamps and healthcare coverage and god knows what else.

I'm playing around a lot with AI, trying to understand it more.

However, I was laid off in April from a terrible marketing and comms job. I think this whole field is going down the tubes because AI can write decent press releases (with a human checkover) and make social posts, etc.

That would also explain why salaries are dropping across the board in the field. And why companies (and state agencies) are so slow to fill positions. I had two good interviews for a state job and we're encroaching on 3 weeks without a decision.

I'm resigning myself to a job fulfilling candy orders for the next few months, but hey, people need their chocolate too.

I've been using Gemini to explore history topics for two years now and it still serves me outrageous porkies like this one. After all the updates and crazy advancements of the past two years, it still makes such a basic mistake.

R104 practically all LLMs, like GPT and Gemini, are notoriously BAD with numbers. Be it dates, basic maths, or formulas.

This is also going on, as described in the tiktok, and all I can say is if you lose or have lost your job due to AI (certain critical tech team, software engineers, project managets to name a few jobs), double your rates/salary requirements, and wait, as patiently as possible, for the time to pounce. It's coming.

Another one re AI eating itself:

Another one by someone who is in charge of managing and implementing AI. Yes, I know tiktok isn't popular here, and I know these are young people, but time to catch up. Legacy and partisan "independent" media is all government or corporate propaganda and tiktok is where the future generations are speaking out, revealing info gleaned from first-hand experience, and spreading info re behind-the-scenes activism. This is the information architecture of this millenium and it won't be the last. Tiktok isn't all silly hot takes or performative bs.