All things AI

[Ditto]

Interesting. What happens when AI gets so smart it works in tandem with the student so that it's impossible to spot? The whole AI thing in society has a long way to go. One thing I hope businesses learn over time is how darn good a well trained human worker can be for customer services and brand reputation. My endlessly repeating experience of corporate services, including with AS (Artificial stupid!) bots, is that we have moved backwards in the last decade, not forwards. In many ways this replicates the sheep like behaviour of organisations uncontrolled export of call centres overseas. They are some good examples of that, but I've seen so many poorly implemented examples, including with a company I worked with - well I just wish companies were smarter and truly focussed on good customer service.

[jmak]

Interesting. What happens when AI gets so smart it works in tandem with the student so that it's impossible to spot? The whole AI thing in society has a long way to go. One thing I hope businesses learn over time is how darn good a well trained human worker can be for customer services and brand reputation. My endlessly repeating experience of corporate services, including with AS (Artificial stupid!) bots, is that we have moved backwards in the last decade, not forwards. In many ways this replicates the sheep like behaviour of organisations uncontrolled export of call centres overseas. They are some good examples of that, but I've seen so many poorly implemented examples, including with a company I worked with - well I just wish companies were smarter and truly focussed on good customer service.

The problem is that everyone says they want excellent customer service, but too few are prepared to pay for it.

[Ditto]

The problem is that everyone says they want excellent customer service, but too few are prepared to pay for it.

 

I don't disagree, but even a basic level of service has disappeared from so many corporations. E.g. My brother need to speak to his Co-op bank. Nearest brach about 20 miles away - opening hours Mon-Fri 9:00 5:00pm. Try to contact by phone - website quotes Lines are open 8am to 6pm Monday to Friday, 9am to 5pm Saturday and Sunday. However, I've got a feeling the weekend service is over stated - regardless, when you ring, you got in to a queue and even after an hour, no answer.

 

I'm trying to access my mortgage with Barclays - online registration isn't working. I've discovered there is a 'local Barclays' nearby, just a few miles awat. It's in a community centre and states "There are no counter or cash services here"! Opening hours are stated as Monday 10.00 am-3.00 pm; Tuesday Closed; Wednesday Closed; Thursday 10.00 am-3.00 pm; Friday 10.00 am-3.00 pm; Saturday Closed; Sunday Closed;

 

Online support? Mon to Fri 9.00am to 5.00pm; Sat to Sun Closed; I did manage to get through once - problem solved - No! You know the appropriate chocolate tea pot phrase. Excuse the language, but FFS, that's insane! Many people are working those hours and it can be hard to make a call and then it can go unanswered.

 

And in case anyone is interested:

 

Co-op bank profits 2023 - £71.4m;

Barclays - well they are struggling compared to the previous year - just £6.5 billion - and in long hand form that's £6,500,000,000.00

[6Foot2]

Link: Copyright and Artificial Intelligence - GOV.UK [Open consultation]

 

Summary

 

This consultation seeks views on how the government can ensure the UK’s legal framework for AI and copyright supports the UK creative industries and AI sector together.

 

This consultation closes at 11:59pm on 25 February 2025

[6Foot2]

Link: Encyclopedia Britannica Is Now an AI Company

 

Once an icon of the 20th century seen as obsolete in the 21st, Encyclopaedia Britannica—now known as just Britannica— is all in on artificial intelligence, and may soon go public at a valuation of nearly $1 billion, according to the New York Times.

 

Until 2012 when printing ended, the company’s books served as the oldest continuously published, English-language encyclopedias in the world, essentially collecting all the world’s knowledge in one place before Google or Wikipedia were a thing. That has helped Britannica pivot into the AI age, where models benefit from access to high-quality, vetted information. More general-purpose models like ChatGPT suffer from hallucinations because they have hoovered up the entire internet, including all the junk and misinformation.

 

While it still offers an online edition of its encyclopedia, as well as the Merriam-Webster dictionary, Britannica’s biggest business today is selling online education software to schools and libraries, the software it hopes to supercharge with AI. That could mean using AI to customize learning plans for individual students. The idea is that students will enjoy learning more when software can help them understand the gaps in their understanding of a topic and stay on it longer. Another education tech company, Brainly, recently announced that answers from its chatbot will link to the exact learning materials (i.e. textbooks) they reference.

 

Britannica’s CEO Jorge Cauz also told the Times about the company’s Britannica AI chatbot, which allows users to ask questions about its vast database of encyclopedic knowledge that it collected over two centuries from vetted academics and editors. The company similarly offers chatbot software for customer service use cases.

 

Britannica told the Times it is expecting revenue to double from two years ago, to $100 million.

 

A company in the space of selling educational books that has seen its fortunes go the opposite direction is Chegg. The company has seen its stock price plummet almost in lock-step with the rise of OpenAI’s ChatGPT, as students canceled their subscriptions to its online knowledge platform.

 

Similar to the rise of Wikipedia before, it seems many people appreciate the access and convenience of ChatGPT, even if they know it cannot be trusted entirely. Chegg has long had an online Q&A platform for homework help where users can pay to submit questions and get answers back. But during the pandemic it saw a flood of new users, threw contractors at answering new questions, and simply could not keep up with ChatGPT. Users have complained that solutions on Chegg are too often wrong, especially when they are submitted by other users rather than professionals.

 

Perhaps the prestigious brand and legacy of Britannica will help it succeed in this new age where chatbots are still apt to return wrong information. It seems like schools are at least willing to pay for access to something they are more confident in.

[sigma]

I see pupils sitting on various platforms that set work - say Sparxmaths with another Window open with Google or whatever, and a lot of copy/pasting of both the question and the answer going on. For some pupils that fail to grasp an understanding of the work they are being set, this is not going to end well.

 

See also essay writers, paraphrasing tools etc.

[Bankesy]

We have recently started using The Key and KeyGPT which has been helpful for polices, interview tasks and interview questions

[elsiegee40]

A colleague of my daughter’s was interviewing over the internet this week and caught the interviewee using ChatGPT to answer the questions.

 

Their answers were rubbish and wooden anyway, but the real giveaway was the reflection of their screen in their glasses.

 

Makes you wonder if prose sections on application forms of any kind are worth bothering with these days. Daughter’s company is now looking at only doing interviews face to face.

[sigma]

https://blog.mahabali.me/educational-technology-2/different-critiques-of-ai-in-education/

 

And

 

https://blogs.lse.ac.uk/highereducation/2025/01/08/five-questions-to-ask-before-adopting-a-new-technology/

[NegativeKillDeath]

A colleague of my daughter’s was interviewing over the internet this week and caught the interviewee using ChatGPT to answer the questions.

 

Their answers were rubbish and wooden anyway, but the real giveaway was the reflection of their screen in their glasses.

 

Makes you wonder if prose sections on application forms of any kind are worth bothering with these days. Daughter’s company is now looking at only doing interviews face to face.

 

I think I could find an hour out of my day to invite them to the face to face second round interview so they can expand on what they said over the internet.

[PotNoodleTech]

If your company ethic has become so work from home centric that your not even bothering to host face to face interviews any more then you kinda deserve all you get (which is potentially to be lumped with someone inadequate for ....... years).

[DalekSec]

A colleague of my daughter’s was interviewing over the internet this week and caught the interviewee using ChatGPT to answer the questions.

 

Their answers were rubbish and wooden anyway, but the real giveaway was the reflection of their screen in their glasses.

 

Makes you wonder if prose sections on application forms of any kind are worth bothering with these days. Daughter’s company is now looking at only doing interviews face to face.

 

I sit in the same cabin as our recruitment team, most of the time they can spot AI, like you said more wooden!

 

Also we now have a disclaimer on jobs that if found that you used AI in the process we won't recruit you.

[pete]

I sit in the same cabin as our recruitment team, most of the time they can spot AI, like you said more wooden!

 

Also we now have a disclaimer on jobs that if found that you used AI in the process we won't recruit you.

 

Do you make it clear you mean inventing untrue things, rather than use from an accessibility standpoint?

 

We're fine with people using it to draft a more coherent/better-written CV, but if they're using it to lie (on the CV or during the interviews), then it's GTFO.

[Ditto]

If your company ethic has become so work from home centric that your not even bothering to host face to face interviews any more then you kinda deserve all you get (which is potentially to be lumped with someone inadequate for ....... years).

 

If I recall correctly, our mod's daughter is working in the US office for a UK company, so it might be a pragmatic arrangement at this stage of the interview. I'm as close to 100% as possible and again it's a pragmatic employer decision as they need the space for those they support. My wife's employer is now forcing staff to do at least 3 days a week in the office, and if you are pro-rata then tough, it's still a 3 day minimum. Quite a few staff were recruited during and post pandemic, so the new policy will lead to resignations. I'd rather work for an employer that can cope with a hybrid environment, that can trust it's staff and managers to do the work they are expected but take in to account individual circumstances and not resort to blanket unconditional policies.

[Andrew_C]

We're fine with people using it to draft a more coherent/better-written CV,

Having just been involved is fixing the sorta-dorta's (v obviously) AI written CV & covering letter, I'm convinced as can be that AI is not the way to go!

[Rob_D]

Having just been involved is fixing the sorta-dorta's (v obviously) AI written CV & covering letter, I'm convinced as can be that AI is not the way to go!

The trick with most of these things is to write it yourself first and then ask AI to suggest improvements. Then it's just suggesting ways using more impactful language or fixing spag errors rather than generating something wholly new which you then have to fix.

[Sephiroth]

The issue with AI in job applications comes when people use it to look more knowledgeable than they actually are. I would personally love to be working with people that understand how to use a powerful tool like AI, not just taking the verbatim result, but correcting it and tailoring it with the view to understand what it is/does/means.

 

It would be like asking someone that's never met you to look up CVs and cover letters for a random job and write you a custom application, expecting it to be flawless. As @Rob_D says, one way to deal with it is to write the skeleton (at least) first, then get AI to offer tweaks.

[dmj]

The trick with most of these things is to write it yourself first and then ask AI to suggest improvements. Then it's just suggesting ways using more impactful language or fixing spag errors rather than generating something wholly new which you then have to fix.

 

+1

 

An itteration on this is to have AI document code, or write unit tests.

Once you've written the code AI is extremely good at reading it and turning those instructions into legible text for your future self to understand what you were doing.

It's not currently so good at doing it the other way around!

[FN-GM]

The trick with most of these things is to write it yourself first and then ask AI to suggest improvements. Then it's just suggesting ways using more impactful language or fixing spag errors rather than generating something wholly new which you then have to fix.

 

This is what I do with global comms at work. Makes life much easier!

[pete]

^ yeah, all of that.

 

I've used it before if I'm converting a document from an adult audience to a "5 year old needs to understand this".

 

Do I still put in front of a couple of test 5 year olds and a teacher of 5 year olds to double-check before it goes live? Of course.

 

Does it (with a reasonable document to work with and a decent prompt) do a fair job of it most of the time? Also yes.

[dmj]

^ yeah, all of that.

 

I've used it before if I'm converting a document from an adult audience to a "5 year old needs to understand this".

 

Do I still put in front of a couple of test 5 year olds and a teacher of 5 year olds to double-check before it goes live? Of course.

 

Does it (with a reasonable document to work with and a decent prompt) do a fair job of it most of the time? Also yes.

 

Anyone tried with "A PE teacher needs to understand this". ??

[pete]

Anyone tried with "A PE teacher needs to understand this". ??

 

"Explain this document using football and other sporting analogies from the UK"

 

Then again, I used Pingu as an example for someone this morning.

[NegativeKillDeath]

Schools have had AI for as long as they've had PE departments.

[Ditto]

We are experimenting with the MS Teams transcribe feature. Our IT support team have once again not shown good knowledge here! I've learnt basic transcription service in include with our main licenses of M365 Apps for business. A small number of staff have the Premium version and this offers more, including one thing our Head of Trustees wants to use, which is AI enabled intelligent recap. From more corporate employers than my own, I have heard positive feedback on this.

 

Q1. Anyone used transcribe and intelligent recap and what is your feedback

Q2. As the premium product cost more pppm, have you looked at other AI apps that integrate with MS Teams - e.g. MeetGeek appears to be one, but I'm sure it is a fast moving feast to know what products are up to scratch?

[6Foot2]

Link: Yes, Minister character is government's new AI assistant - BBC News

 

Government workers will soon be given access to a set of tools powered by artificial intelligence (AI), named after a scheming parliamentary official from the classic sitcom Yes, Minister.

 

The government says the assistants - called Humphrey - will "speed up the work of civil servants" and save money by replacing cash that would have been spent on consultants.

 

But the decision to name the AI after Sir Humphrey Appleby, a character described as "devious and controlling", has raised eyebrows.

 

Tim Flagg, chief operating officer of trade body UKAI, said the name risks "undermining" the government's mission to embrace the tech.

 

Science and technology secretary Peter Kyle will announce more digital tools later on Tuesday, including two apps which will store government documents, including digital driving licenses.

 

The announcement is part of the government's overhaul of digital services and comes after their AI Opportunities Action Plan announced last week.

 

"Humphrey for me is a name which is very associated with the Machiavellian character from Yes, Minister," says Mr Flagg from UKAI, which represents the AI sector.

 

"That immediately makes people who aren't in that central Whitehall office think that this is something which is not going to be empowering and not going to be helping them."

 

Most of the tools in the Humphrey suite are generative AI models - in this case, technology which takes large amounts of information and summarises it in a more digestible format - to be used by the civil service.

 

Among them is Consult, external, which summarises people's responses to public calls for information.

 

The government says this is currently done by expensive external consultants who bill the taxpayer "around £100,000 every time."

 

Parlex, external, which the government says helps policymakers search through previous parliamentary debates on a certain topic, is described by The Times, external as "designed to avoid catastrophic political rows by predicting how MPs will respond".

 

Other changes announced include more efficient data sharing between departments.

 

"I think the government is doing the right thing," says Mr Flagg.

 

"They do have some good developers - I have every confidence they are going to be creating a great product."